qurator-spk/eynollah
1
0
Fork 0
mirror of https://github.com/qurator-spk/eynollah.git synced 2025-10-31 01:34:17 +01:00
Code Issues Projects Releases Wiki Activity

flow from directory

Browse source
This commit is contained in:
vahid 2022-03-29 06:55:19 -04:00
parent cf5ef8f5ae
commit c606391c31
4 changed files with 767 additions and 437 deletions
Download patch file Download diff file Expand all files Collapse all files

14
qurator/eynollah/cli.py
View file

@ -10,7 +10,6 @@ from qurator.eynollah.eynollah import Eynollah
"-i", "-i",
help="image filename", help="image filename",
type=click.Path(exists=True, dir_okay=False), type=click.Path(exists=True, dir_okay=False),
required=True,
) )
@click.option( @click.option(
"--out", "--out",
@ -19,6 +18,12 @@ from qurator.eynollah.eynollah import Eynollah
type=click.Path(exists=True, file_okay=False), type=click.Path(exists=True, file_okay=False),
required=True, required=True,
) )
@click.option(
"--dir_in",
"-di",
help="directory of images",
type=click.Path(exists=True, file_okay=False),
)
@click.option( @click.option(
"--model", "--model",
"-m", "-m",
@ -112,6 +117,7 @@ from qurator.eynollah.eynollah import Eynollah
def main( def main(
image, image,
out, out,
dir_in,
model, model,
save_images, save_images,
save_layout, save_layout,
@ -140,6 +146,7 @@ def main(
eynollah = Eynollah( eynollah = Eynollah(
image_filename=image, image_filename=image,
dir_out=out, dir_out=out,
dir_in=dir_in,
dir_models=model, dir_models=model,
dir_of_cropped_images=save_images, dir_of_cropped_images=save_images,
dir_of_layout=save_layout, dir_of_layout=save_layout,
@ -155,8 +162,9 @@ def main(
headers_off=headers_off, headers_off=headers_off,
light_version=light_version, light_version=light_version,
) )
pcgts = eynollah.run() eynollah.run()
eynollah.writer.write_pagexml(pcgts) #pcgts = eynollah.run()
##eynollah.writer.write_pagexml(pcgts)
if __name__ == "__main__": if __name__ == "__main__":
main() main()

323
qurator/eynollah/eynollah.py
View file

@ -29,6 +29,7 @@ warnings.filterwarnings("ignore")
from scipy.signal import find_peaks from scipy.signal import find_peaks
import matplotlib.pyplot as plt import matplotlib.pyplot as plt
from scipy.ndimage import gaussian_filter1d from scipy.ndimage import gaussian_filter1d
from keras.backend import set_session
from .utils.contour import ( from .utils.contour import (
filter_contours_area_of_image, filter_contours_area_of_image,
@ -38,6 +39,7 @@ from .utils.contour import (
find_features_of_contours, find_features_of_contours,
get_text_region_boxes_by_given_contours, get_text_region_boxes_by_given_contours,
get_textregion_contours_in_org_image, get_textregion_contours_in_org_image,
get_textregion_contours_in_org_image_light,
return_contours_of_image, return_contours_of_image,
return_contours_of_interested_region, return_contours_of_interested_region,
return_contours_of_interested_region_by_min_size, return_contours_of_interested_region_by_min_size,
@ -65,6 +67,7 @@ from .utils import (
put_drop_out_from_only_drop_model, put_drop_out_from_only_drop_model,
putt_bb_of_drop_capitals_of_model_in_patches_in_layout, putt_bb_of_drop_capitals_of_model_in_patches_in_layout,
check_any_text_region_in_model_one_is_main_or_header, check_any_text_region_in_model_one_is_main_or_header,
check_any_text_region_in_model_one_is_main_or_header_light,
small_textlines_to_parent_adherence2, small_textlines_to_parent_adherence2,
order_of_regions, order_of_regions,
find_number_of_columns_in_document, find_number_of_columns_in_document,
@ -84,10 +87,11 @@ class Eynollah:
def __init__( def __init__(
self, self,
dir_models, dir_models,
image_filename, image_filename=None,
image_pil=None, image_pil=None,
image_filename_stem=None, image_filename_stem=None,
dir_out=None, dir_out=None,
dir_in=None,
dir_of_cropped_images=None, dir_of_cropped_images=None,
dir_of_layout=None, dir_of_layout=None,
dir_of_deskewed=None, dir_of_deskewed=None,
@ -105,6 +109,7 @@ class Eynollah:
logger=None, logger=None,
pcgts=None, pcgts=None,
): ):
if not dir_in:
if image_pil: if image_pil:
self._imgs = self._cache_images(image_pil=image_pil) self._imgs = self._cache_images(image_pil=image_pil)
else: else:
@ -113,6 +118,7 @@ class Eynollah:
self.dpi = override_dpi self.dpi = override_dpi
self.image_filename = image_filename self.image_filename = image_filename
self.dir_out = dir_out self.dir_out = dir_out
self.dir_in = dir_in
self.allow_enhancement = allow_enhancement self.allow_enhancement = allow_enhancement
self.curved_line = curved_line self.curved_line = curved_line
self.full_layout = full_layout self.full_layout = full_layout
@ -121,6 +127,7 @@ class Eynollah:
self.allow_scaling = allow_scaling self.allow_scaling = allow_scaling
self.headers_off = headers_off self.headers_off = headers_off
self.light_version = light_version self.light_version = light_version
self.pcgts = pcgts
self.plotter = None if not enable_plotting else EynollahPlotter( self.plotter = None if not enable_plotting else EynollahPlotter(
dir_out=self.dir_out, dir_out=self.dir_out,
dir_of_all=dir_of_all, dir_of_all=dir_of_all,
@ -128,6 +135,7 @@ class Eynollah:
dir_of_cropped_images=dir_of_cropped_images, dir_of_cropped_images=dir_of_cropped_images,
dir_of_layout=dir_of_layout, dir_of_layout=dir_of_layout,
image_filename_stem=Path(Path(image_filename).name).stem) image_filename_stem=Path(Path(image_filename).name).stem)
if not dir_in:
self.writer = EynollahXmlWriter( self.writer = EynollahXmlWriter(
dir_out=self.dir_out, dir_out=self.dir_out,
image_filename=self.image_filename, image_filename=self.image_filename,
@ -149,6 +157,41 @@ class Eynollah:
self.model_textline_dir = dir_models + "/model_textline_newspapers.h5" self.model_textline_dir = dir_models + "/model_textline_newspapers.h5"
self.model_tables = dir_models + "/model_tables_ens_mixed_new_2.h5" self.model_tables = dir_models + "/model_tables_ens_mixed_new_2.h5"
if dir_in and light_version:
config = tf.compat.v1.ConfigProto()
config.gpu_options.allow_growth = True
session = tf.compat.v1.Session(config=config)
set_session(session)
self.model_page = self.our_load_model(self.model_page_dir)
self.model_classifier = self.our_load_model(self.model_dir_of_col_classifier)
self.model_bin = self.our_load_model(self.model_dir_of_binarization)
self.model_textline = self.our_load_model(self.model_textline_dir)
self.model_region = self.our_load_model(self.model_region_dir_p_ens_light)
self.model_region_fl_np = self.our_load_model(self.model_region_dir_fully_np)
self.model_region_fl = self.our_load_model(self.model_region_dir_fully)
self.ls_imgs = os.listdir(self.dir_in)
if dir_in and not light_version:
config = tf.compat.v1.ConfigProto()
config.gpu_options.allow_growth = True
session = tf.compat.v1.Session(config=config)
set_session(session)
self.model_page = self.our_load_model(self.model_page_dir)
self.model_classifier = self.our_load_model(self.model_dir_of_col_classifier)
self.model_bin = self.our_load_model(self.model_dir_of_binarization)
self.model_textline = self.our_load_model(self.model_textline_dir)
self.model_region = self.our_load_model(self.model_region_dir_p_ens)
self.model_region_p2 = self.our_load_model(self.model_region_dir_p2)
self.model_region_fl_np = self.our_load_model(self.model_region_dir_fully_np)
self.model_region_fl = self.our_load_model(self.model_region_dir_fully)
self.model_enhancement = self.our_load_model(self.model_dir_of_enhancement)
self.ls_imgs = os.listdir(self.dir_in)
def _cache_images(self, image_filename=None, image_pil=None): def _cache_images(self, image_filename=None, image_pil=None):
ret = {} ret = {}
if image_filename: if image_filename:
@ -161,7 +204,15 @@ class Eynollah:
for prefix in ('', '_grayscale'): for prefix in ('', '_grayscale'):
ret[f'img{prefix}_uint8'] = ret[f'img{prefix}'].astype(np.uint8) ret[f'img{prefix}_uint8'] = ret[f'img{prefix}'].astype(np.uint8)
return ret return ret
def reset_file_name_dir(self, image_filename):
self._imgs = self._cache_images(image_filename=image_filename)
self.image_filename = image_filename
self.writer = EynollahXmlWriter(
dir_out=self.dir_out,
image_filename=self.image_filename,
curved_line=self.curved_line,
pcgts=self.pcgts)
def imread(self, grayscale=False, uint8=True): def imread(self, grayscale=False, uint8=True):
key = 'img' key = 'img'
if grayscale: if grayscale:
@ -335,6 +386,7 @@ class Eynollah:
img = self.imread() img = self.imread()
_, page_coord = self.early_page_for_num_of_column_classification(img) _, page_coord = self.early_page_for_num_of_column_classification(img)
if not self.dir_in:
model_num_classifier, session_col_classifier = self.start_new_session_and_model(self.model_dir_of_col_classifier) model_num_classifier, session_col_classifier = self.start_new_session_and_model(self.model_dir_of_col_classifier)
if self.input_binary: if self.input_binary:
img_in = np.copy(img) img_in = np.copy(img)
@ -357,17 +409,18 @@ class Eynollah:
img_in[0, :, :, 0] = img_1ch[:, :] img_in[0, :, :, 0] = img_1ch[:, :]
img_in[0, :, :, 1] = img_1ch[:, :] img_in[0, :, :, 1] = img_1ch[:, :]
img_in[0, :, :, 2] = img_1ch[:, :] img_in[0, :, :, 2] = img_1ch[:, :]
if not self.dir_in:
label_p_pred = model_num_classifier.predict(img_in) label_p_pred = model_num_classifier.predict(img_in)
else:
label_p_pred = self.model_classifier.predict(img_in)
num_col = np.argmax(label_p_pred[0]) + 1 num_col = np.argmax(label_p_pred[0]) + 1
self.logger.info("Found %s columns (%s)", num_col, label_p_pred) self.logger.info("Found %s columns (%s)", num_col, label_p_pred)
if not self.dir_in:
session_col_classifier.close() session_col_classifier.close()
del model_num_classifier del model_num_classifier
del session_col_classifier del session_col_classifier
K.clear_session() K.clear_session()
gc.collect() gc.collect()
@ -383,13 +436,14 @@ class Eynollah:
def resize_and_enhance_image_with_column_classifier(self,light_version): def resize_and_enhance_image_with_column_classifier(self,light_version):
self.logger.debug("enter resize_and_enhance_image_with_column_classifier") self.logger.debug("enter resize_and_enhance_image_with_column_classifier")
if light_version:
dpi = 300
else:
dpi = self.dpi dpi = self.dpi
self.logger.info("Detected %s DPI", dpi) self.logger.info("Detected %s DPI", dpi)
if self.input_binary: if self.input_binary:
img = self.imread() img = self.imread()
if self.dir_in:
prediction_bin = self.do_prediction(True, img, self.model_bin)
else:
model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization) model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization)
prediction_bin = self.do_prediction(True, img, model_bin) prediction_bin = self.do_prediction(True, img, model_bin)
@ -399,6 +453,7 @@ class Eynollah:
prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2)
if not self.dir_in:
session_bin.close() session_bin.close()
del model_bin del model_bin
del session_bin del session_bin
@ -412,6 +467,7 @@ class Eynollah:
img_bin = None img_bin = None
_, page_coord = self.early_page_for_num_of_column_classification(img_bin) _, page_coord = self.early_page_for_num_of_column_classification(img_bin)
if not self.dir_in:
model_num_classifier, session_col_classifier = self.start_new_session_and_model(self.model_dir_of_col_classifier) model_num_classifier, session_col_classifier = self.start_new_session_and_model(self.model_dir_of_col_classifier)
if self.input_binary: if self.input_binary:
@ -433,22 +489,28 @@ class Eynollah:
img_in[0, :, :, 2] = img_1ch[:, :] img_in[0, :, :, 2] = img_1ch[:, :]
if self.dir_in:
label_p_pred = self.model_classifier.predict(img_in)
else:
label_p_pred = model_num_classifier.predict(img_in) label_p_pred = model_num_classifier.predict(img_in)
num_col = np.argmax(label_p_pred[0]) + 1 num_col = np.argmax(label_p_pred[0]) + 1
self.logger.info("Found %s columns (%s)", num_col, label_p_pred) self.logger.info("Found %s columns (%s)", num_col, label_p_pred)
if not self.dir_in:
session_col_classifier.close() session_col_classifier.close()
K.clear_session() K.clear_session()
if dpi < DPI_THRESHOLD: if dpi < DPI_THRESHOLD:
img_new, num_column_is_classified = self.calculate_width_height_by_columns(img, num_col, width_early, label_p_pred) img_new, num_column_is_classified = self.calculate_width_height_by_columns(img, num_col, width_early, label_p_pred)
if light_version:
image_res = np.copy(img_new)
else:
image_res = self.predict_enhancement(img_new) image_res = self.predict_enhancement(img_new)
is_image_enhanced = True is_image_enhanced = True
else: else:
num_column_is_classified = True num_column_is_classified = True
image_res = np.copy(img) image_res = np.copy(img)
is_image_enhanced = False is_image_enhanced = False
if not self.dir_in:
session_col_classifier.close() session_col_classifier.close()
@ -595,48 +657,48 @@ class Eynollah:
if i == 0 and j == 0: if i == 0 and j == 0:
seg_color = seg_color[0 : seg_color.shape[0] - margin, 0 : seg_color.shape[1] - margin, :] seg_color = seg_color[0 : seg_color.shape[0] - margin, 0 : seg_color.shape[1] - margin, :]
seg = seg[0 : seg.shape[0] - margin, 0 : seg.shape[1] - margin] #seg = seg[0 : seg.shape[0] - margin, 0 : seg.shape[1] - margin]
mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + 0 : index_x_u - margin] = seg #mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + 0 : index_x_u - margin] = seg
prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + 0 : index_x_u - margin, :] = seg_color prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + 0 : index_x_u - margin, :] = seg_color
elif i == nxf - 1 and j == nyf - 1: elif i == nxf - 1 and j == nyf - 1:
seg_color = seg_color[margin : seg_color.shape[0] - 0, margin : seg_color.shape[1] - 0, :] seg_color = seg_color[margin : seg_color.shape[0] - 0, margin : seg_color.shape[1] - 0, :]
seg = seg[margin : seg.shape[0] - 0, margin : seg.shape[1] - 0] #seg = seg[margin : seg.shape[0] - 0, margin : seg.shape[1] - 0]
mask_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - 0] = seg #mask_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - 0] = seg
prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - 0, :] = seg_color prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - 0, :] = seg_color
elif i == 0 and j == nyf - 1: elif i == 0 and j == nyf - 1:
seg_color = seg_color[margin : seg_color.shape[0] - 0, 0 : seg_color.shape[1] - margin, :] seg_color = seg_color[margin : seg_color.shape[0] - 0, 0 : seg_color.shape[1] - margin, :]
seg = seg[margin : seg.shape[0] - 0, 0 : seg.shape[1] - margin] #seg = seg[margin : seg.shape[0] - 0, 0 : seg.shape[1] - margin]
mask_true[index_y_d + margin : index_y_u - 0, index_x_d + 0 : index_x_u - margin] = seg #mask_true[index_y_d + margin : index_y_u - 0, index_x_d + 0 : index_x_u - margin] = seg
prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + 0 : index_x_u - margin, :] = seg_color prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + 0 : index_x_u - margin, :] = seg_color
elif i == nxf - 1 and j == 0: elif i == nxf - 1 and j == 0:
seg_color = seg_color[0 : seg_color.shape[0] - margin, margin : seg_color.shape[1] - 0, :] seg_color = seg_color[0 : seg_color.shape[0] - margin, margin : seg_color.shape[1] - 0, :]
seg = seg[0 : seg.shape[0] - margin, margin : seg.shape[1] - 0] #seg = seg[0 : seg.shape[0] - margin, margin : seg.shape[1] - 0]
mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - 0] = seg #mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - 0] = seg
prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - 0, :] = seg_color prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - 0, :] = seg_color
elif i == 0 and j != 0 and j != nyf - 1: elif i == 0 and j != 0 and j != nyf - 1:
seg_color = seg_color[margin : seg_color.shape[0] - margin, 0 : seg_color.shape[1] - margin, :] seg_color = seg_color[margin : seg_color.shape[0] - margin, 0 : seg_color.shape[1] - margin, :]
seg = seg[margin : seg.shape[0] - margin, 0 : seg.shape[1] - margin] #seg = seg[margin : seg.shape[0] - margin, 0 : seg.shape[1] - margin]
mask_true[index_y_d + margin : index_y_u - margin, index_x_d + 0 : index_x_u - margin] = seg #mask_true[index_y_d + margin : index_y_u - margin, index_x_d + 0 : index_x_u - margin] = seg
prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + 0 : index_x_u - margin, :] = seg_color prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + 0 : index_x_u - margin, :] = seg_color
elif i == nxf - 1 and j != 0 and j != nyf - 1: elif i == nxf - 1 and j != 0 and j != nyf - 1:
seg_color = seg_color[margin : seg_color.shape[0] - margin, margin : seg_color.shape[1] - 0, :] seg_color = seg_color[margin : seg_color.shape[0] - margin, margin : seg_color.shape[1] - 0, :]
seg = seg[margin : seg.shape[0] - margin, margin : seg.shape[1] - 0] #seg = seg[margin : seg.shape[0] - margin, margin : seg.shape[1] - 0]
mask_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - 0] = seg #mask_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - 0] = seg
prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - 0, :] = seg_color prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - 0, :] = seg_color
elif i != 0 and i != nxf - 1 and j == 0: elif i != 0 and i != nxf - 1 and j == 0:
seg_color = seg_color[0 : seg_color.shape[0] - margin, margin : seg_color.shape[1] - margin, :] seg_color = seg_color[0 : seg_color.shape[0] - margin, margin : seg_color.shape[1] - margin, :]
seg = seg[0 : seg.shape[0] - margin, margin : seg.shape[1] - margin] #seg = seg[0 : seg.shape[0] - margin, margin : seg.shape[1] - margin]
mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - margin] = seg #mask_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - margin] = seg
prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color prediction_true[index_y_d + 0 : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color
elif i != 0 and i != nxf - 1 and j == nyf - 1: elif i != 0 and i != nxf - 1 and j == nyf - 1:
seg_color = seg_color[margin : seg_color.shape[0] - 0, margin : seg_color.shape[1] - margin, :] seg_color = seg_color[margin : seg_color.shape[0] - 0, margin : seg_color.shape[1] - margin, :]
seg = seg[margin : seg.shape[0] - 0, margin : seg.shape[1] - margin] #seg = seg[margin : seg.shape[0] - 0, margin : seg.shape[1] - margin]
mask_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - margin] = seg #mask_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - margin] = seg
prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - margin, :] = seg_color prediction_true[index_y_d + margin : index_y_u - 0, index_x_d + margin : index_x_u - margin, :] = seg_color
else: else:
seg_color = seg_color[margin : seg_color.shape[0] - margin, margin : seg_color.shape[1] - margin, :] seg_color = seg_color[margin : seg_color.shape[0] - margin, margin : seg_color.shape[1] - margin, :]
seg = seg[margin : seg.shape[0] - margin, margin : seg.shape[1] - margin] #seg = seg[margin : seg.shape[0] - margin, margin : seg.shape[1] - margin]
mask_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin] = seg #mask_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin] = seg
prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color prediction_true[index_y_d + margin : index_y_u - margin, index_x_d + margin : index_x_u - margin, :] = seg_color
prediction_true = prediction_true.astype(np.uint8) prediction_true = prediction_true.astype(np.uint8)
@ -817,9 +879,12 @@ class Eynollah:
img = img.astype(np.uint8) img = img.astype(np.uint8)
else: else:
img = self.imread() img = self.imread()
if not self.dir_in:
model_page, session_page = self.start_new_session_and_model(self.model_page_dir) model_page, session_page = self.start_new_session_and_model(self.model_page_dir)
img = cv2.GaussianBlur(img, (5, 5), 0) img = cv2.GaussianBlur(img, (5, 5), 0)
if self.dir_in:
img_page_prediction = self.do_prediction(False, img, self.model_page)
else:
img_page_prediction = self.do_prediction(False, img, model_page) img_page_prediction = self.do_prediction(False, img, model_page)
imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY) imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY)
@ -834,20 +899,25 @@ class Eynollah:
else: else:
box = [0, 0, img.shape[1], img.shape[0]] box = [0, 0, img.shape[1], img.shape[0]]
croped_page, page_coord = crop_image_inside_box(box, img) croped_page, page_coord = crop_image_inside_box(box, img)
if not self.dir_in:
session_page.close() session_page.close()
del model_page del model_page
del session_page del session_page
gc.collect()
K.clear_session() K.clear_session()
gc.collect()
self.logger.debug("exit early_page_for_num_of_column_classification") self.logger.debug("exit early_page_for_num_of_column_classification")
return croped_page, page_coord return croped_page, page_coord
def extract_page(self): def extract_page(self):
self.logger.debug("enter extract_page") self.logger.debug("enter extract_page")
cont_page = [] cont_page = []
if not self.dir_in:
model_page, session_page = self.start_new_session_and_model(self.model_page_dir) model_page, session_page = self.start_new_session_and_model(self.model_page_dir)
img = cv2.GaussianBlur(self.image, (5, 5), 0) img = cv2.GaussianBlur(self.image, (5, 5), 0)
if not self.dir_in:
img_page_prediction = self.do_prediction(False, img, model_page) img_page_prediction = self.do_prediction(False, img, model_page)
else:
img_page_prediction = self.do_prediction(False, img, self.model_page)
imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY) imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY)
_, thresh = cv2.threshold(imgray, 0, 255, 0) _, thresh = cv2.threshold(imgray, 0, 255, 0)
thresh = cv2.dilate(thresh, KERNEL, iterations=3) thresh = cv2.dilate(thresh, KERNEL, iterations=3)
@ -873,11 +943,12 @@ class Eynollah:
box = [0, 0, img.shape[1], img.shape[0]] box = [0, 0, img.shape[1], img.shape[0]]
croped_page, page_coord = crop_image_inside_box(box, self.image) croped_page, page_coord = crop_image_inside_box(box, self.image)
cont_page.append(np.array([[page_coord[2], page_coord[0]], [page_coord[3], page_coord[0]], [page_coord[3], page_coord[1]], [page_coord[2], page_coord[1]]])) cont_page.append(np.array([[page_coord[2], page_coord[0]], [page_coord[3], page_coord[0]], [page_coord[3], page_coord[1]], [page_coord[2], page_coord[1]]]))
if not self.dir_in:
session_page.close() session_page.close()
del model_page del model_page
del session_page del session_page
gc.collect()
K.clear_session() K.clear_session()
gc.collect()
self.logger.debug("exit extract_page") self.logger.debug("exit extract_page")
return croped_page, page_coord, cont_page return croped_page, page_coord, cont_page
@ -888,9 +959,13 @@ class Eynollah:
img = img.astype(np.uint8) img = img.astype(np.uint8)
else: else:
img = self.imread() img = self.imread()
if not self.dir_in:
model_page, session_page = self.start_new_session_and_model(self.model_page_dir) model_page, session_page = self.start_new_session_and_model(self.model_page_dir)
img = cv2.GaussianBlur(img, (5, 5), 0) img = cv2.GaussianBlur(img, (5, 5), 0)
if self.dir_in:
img_page_prediction = self.do_prediction(False, img, self.model_page)
else:
img_page_prediction = self.do_prediction(False, img, model_page) img_page_prediction = self.do_prediction(False, img, model_page)
imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY) imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY)
@ -905,20 +980,28 @@ class Eynollah:
else: else:
box = [0, 0, img.shape[1], img.shape[0]] box = [0, 0, img.shape[1], img.shape[0]]
croped_page, page_coord = crop_image_inside_box(box, img) croped_page, page_coord = crop_image_inside_box(box, img)
if not self.dir_in:
session_page.close() session_page.close()
del model_page del model_page
del session_page del session_page
gc.collect()
K.clear_session() K.clear_session()
gc.collect()
self.logger.debug("exit early_page_for_num_of_column_classification") self.logger.debug("exit early_page_for_num_of_column_classification")
return croped_page, page_coord return croped_page, page_coord
def extract_page(self): def extract_page(self):
self.logger.debug("enter extract_page") self.logger.debug("enter extract_page")
cont_page = [] cont_page = []
if not self.dir_in:
model_page, session_page = self.start_new_session_and_model(self.model_page_dir) model_page, session_page = self.start_new_session_and_model(self.model_page_dir)
img = cv2.GaussianBlur(self.image, (5, 5), 0) img = cv2.GaussianBlur(self.image, (5, 5), 0)
if not self.dir_in:
img_page_prediction = self.do_prediction(False, img, model_page) img_page_prediction = self.do_prediction(False, img, model_page)
else:
img_page_prediction = self.do_prediction(False, img, self.model_page)
imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY) imgray = cv2.cvtColor(img_page_prediction, cv2.COLOR_BGR2GRAY)
_, thresh = cv2.threshold(imgray, 0, 255, 0) _, thresh = cv2.threshold(imgray, 0, 255, 0)
thresh = cv2.dilate(thresh, KERNEL, iterations=3) thresh = cv2.dilate(thresh, KERNEL, iterations=3)
@ -944,11 +1027,12 @@ class Eynollah:
box = [0, 0, img.shape[1], img.shape[0]] box = [0, 0, img.shape[1], img.shape[0]]
croped_page, page_coord = crop_image_inside_box(box, self.image) croped_page, page_coord = crop_image_inside_box(box, self.image)
cont_page.append(np.array([[page_coord[2], page_coord[0]], [page_coord[3], page_coord[0]], [page_coord[3], page_coord[1]], [page_coord[2], page_coord[1]]])) cont_page.append(np.array([[page_coord[2], page_coord[0]], [page_coord[3], page_coord[0]], [page_coord[3], page_coord[1]], [page_coord[2], page_coord[1]]]))
if not self.dir_in:
session_page.close() session_page.close()
del model_page del model_page
del session_page del session_page
gc.collect()
K.clear_session() K.clear_session()
gc.collect()
self.logger.debug("exit extract_page") self.logger.debug("exit extract_page")
return croped_page, page_coord, cont_page return croped_page, page_coord, cont_page
@ -956,8 +1040,10 @@ class Eynollah:
self.logger.debug("enter extract_text_regions") self.logger.debug("enter extract_text_regions")
img_height_h = img.shape[0] img_height_h = img.shape[0]
img_width_h = img.shape[1] img_width_h = img.shape[1]
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_fully if patches else self.model_region_dir_fully_np) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_fully if patches else self.model_region_dir_fully_np)
else:
model_region = self.model_region_fl if patches else self.model_region_fl_np
if not patches: if not patches:
img = otsu_copy_binary(img) img = otsu_copy_binary(img)
@ -1044,6 +1130,7 @@ class Eynollah:
prediction_regions = self.do_prediction(patches, img, model_region, marginal_of_patch_percent) prediction_regions = self.do_prediction(patches, img, model_region, marginal_of_patch_percent)
prediction_regions = resize_image(prediction_regions, img_height_h, img_width_h) prediction_regions = resize_image(prediction_regions, img_height_h, img_width_h)
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
@ -1422,18 +1509,25 @@ class Eynollah:
def textline_contours(self, img, patches, scaler_h, scaler_w): def textline_contours(self, img, patches, scaler_h, scaler_w):
self.logger.debug('enter textline_contours') self.logger.debug('enter textline_contours')
if not self.dir_in:
model_textline, session_textline = self.start_new_session_and_model(self.model_textline_dir if patches else self.model_textline_dir_np) model_textline, session_textline = self.start_new_session_and_model(self.model_textline_dir if patches else self.model_textline_dir_np)
img = img.astype(np.uint8) img = img.astype(np.uint8)
img_org = np.copy(img) img_org = np.copy(img)
img_h = img_org.shape[0] img_h = img_org.shape[0]
img_w = img_org.shape[1] img_w = img_org.shape[1]
img = resize_image(img_org, int(img_org.shape[0] * scaler_h), int(img_org.shape[1] * scaler_w)) img = resize_image(img_org, int(img_org.shape[0] * scaler_h), int(img_org.shape[1] * scaler_w))
if not self.dir_in:
prediction_textline = self.do_prediction(patches, img, model_textline) prediction_textline = self.do_prediction(patches, img, model_textline)
else:
prediction_textline = self.do_prediction(patches, img, self.model_textline)
prediction_textline = resize_image(prediction_textline, img_h, img_w) prediction_textline = resize_image(prediction_textline, img_h, img_w)
if not self.dir_in:
prediction_textline_longshot = self.do_prediction(False, img, model_textline) prediction_textline_longshot = self.do_prediction(False, img, model_textline)
else:
prediction_textline_longshot = self.do_prediction(False, img, self.model_textline)
prediction_textline_longshot_true_size = resize_image(prediction_textline_longshot, img_h, img_w) prediction_textline_longshot_true_size = resize_image(prediction_textline_longshot, img_h, img_w)
if not self.dir_in:
session_textline.close() session_textline.close()
@ -1508,17 +1602,17 @@ class Eynollah:
##img_resized = resize_image(img_bin,img_height_h, img_width_h ) ##img_resized = resize_image(img_bin,img_height_h, img_width_h )
img_resized = resize_image(img,img_h_new, img_w_new ) img_resized = resize_image(img,img_h_new, img_w_new )
tbin = time.time() if not self.dir_in:
model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization) model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization)
print("time bin session", time.time()-tbin)
prediction_bin = self.do_prediction(True, img_resized, model_bin) prediction_bin = self.do_prediction(True, img_resized, model_bin)
print("time bin all ", time.time()-tbin) else:
prediction_bin = self.do_prediction(True, img_resized, self.model_bin)
prediction_bin=prediction_bin[:,:,0] prediction_bin=prediction_bin[:,:,0]
prediction_bin = (prediction_bin[:,:]==0)*1 prediction_bin = (prediction_bin[:,:]==0)*1
prediction_bin = prediction_bin*255 prediction_bin = prediction_bin*255
prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2)
if not self.dir_in:
session_bin.close() session_bin.close()
del model_bin del model_bin
del session_bin del session_bin
@ -1530,17 +1624,13 @@ class Eynollah:
tline = time.time()
textline_mask_tot_ea = self.run_textline(img_bin) textline_mask_tot_ea = self.run_textline(img_bin)
print("time line all ", time.time()-tline)
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens_light) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens_light)
#plt.imshow(img_bin)
#plt.show()
prediction_regions_org = self.do_prediction_new_concept(True, img_bin, model_region) prediction_regions_org = self.do_prediction_new_concept(True, img_bin, model_region)
else:
prediction_regions_org = self.do_prediction_new_concept(True, img_bin, self.model_region)
#plt.imshow(prediction_regions_org[:,:,0]) #plt.imshow(prediction_regions_org[:,:,0])
#plt.show() #plt.show()
@ -1564,7 +1654,6 @@ class Eynollah:
polygons_of_only_lines = return_contours_of_interested_region(mask_lines_only,1,0.00001) polygons_of_only_lines = return_contours_of_interested_region(mask_lines_only,1,0.00001)
text_regions_p_true = np.zeros(prediction_regions_org.shape) text_regions_p_true = np.zeros(prediction_regions_org.shape)
text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_lines, color=(3,3,3)) text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_lines, color=(3,3,3))
@ -1574,6 +1663,7 @@ class Eynollah:
text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1)) text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1))
#erosion_hurts = True #erosion_hurts = True
if not self.dir_in:
K.clear_session() K.clear_session()
return text_regions_p_true, erosion_hurts, polygons_lines_xml, textline_mask_tot_ea return text_regions_p_true, erosion_hurts, polygons_lines_xml, textline_mask_tot_ea
@ -1584,14 +1674,17 @@ class Eynollah:
img_height_h = img_org.shape[0] img_height_h = img_org.shape[0]
img_width_h = img_org.shape[1] img_width_h = img_org.shape[1]
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens)
ratio_y=1.3 ratio_y=1.3
ratio_x=1 ratio_x=1
img = resize_image(img_org, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x)) img = resize_image(img_org, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x))
if not self.dir_in:
prediction_regions_org_y = self.do_prediction(True, img, model_region) prediction_regions_org_y = self.do_prediction(True, img, model_region)
else:
prediction_regions_org_y = self.do_prediction(True, img, self.model_region)
prediction_regions_org_y = resize_image(prediction_regions_org_y, img_height_h, img_width_h ) prediction_regions_org_y = resize_image(prediction_regions_org_y, img_height_h, img_width_h )
#plt.imshow(prediction_regions_org_y[:,:,0]) #plt.imshow(prediction_regions_org_y[:,:,0])
@ -1610,6 +1703,9 @@ class Eynollah:
img = resize_image(img_org, int(img_org.shape[0]), int(img_org.shape[1]*(1.2 if is_image_enhanced else 1))) img = resize_image(img_org, int(img_org.shape[0]), int(img_org.shape[1]*(1.2 if is_image_enhanced else 1)))
if self.dir_in:
prediction_regions_org = self.do_prediction(True, img, self.model_region)
else:
prediction_regions_org = self.do_prediction(True, img, model_region) prediction_regions_org = self.do_prediction(True, img, model_region)
prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h ) prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h )
@ -1618,17 +1714,23 @@ class Eynollah:
prediction_regions_org=prediction_regions_org[:,:,0] prediction_regions_org=prediction_regions_org[:,:,0]
prediction_regions_org[(prediction_regions_org[:,:]==1) & (mask_zeros_y[:,:]==1)]=0 prediction_regions_org[(prediction_regions_org[:,:]==1) & (mask_zeros_y[:,:]==1)]=0
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
gc.collect() gc.collect()
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p2) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p2)
img = resize_image(img_org, int(img_org.shape[0]), int(img_org.shape[1])) img = resize_image(img_org, int(img_org.shape[0]), int(img_org.shape[1]))
if self.dir_in:
prediction_regions_org2 = self.do_prediction(True, img, self.model_region_p2, 0.2)
else:
prediction_regions_org2 = self.do_prediction(True, img, model_region, 0.2) prediction_regions_org2 = self.do_prediction(True, img, model_region, 0.2)
prediction_regions_org2=resize_image(prediction_regions_org2, img_height_h, img_width_h ) prediction_regions_org2=resize_image(prediction_regions_org2, img_height_h, img_width_h )
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
@ -1663,8 +1765,11 @@ class Eynollah:
if self.input_binary: if self.input_binary:
prediction_bin = np.copy(img_org) prediction_bin = np.copy(img_org)
else: else:
if not self.dir_in:
model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization) model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization)
prediction_bin = self.do_prediction(True, img_org, model_bin) prediction_bin = self.do_prediction(True, img_org, model_bin)
else:
prediction_bin = self.do_prediction(True, img_org, self.model_bin)
prediction_bin = resize_image(prediction_bin, img_height_h, img_width_h ) prediction_bin = resize_image(prediction_bin, img_height_h, img_width_h )
prediction_bin=prediction_bin[:,:,0] prediction_bin=prediction_bin[:,:,0]
@ -1673,13 +1778,14 @@ class Eynollah:
prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2)
if not self.dir_in:
session_bin.close() session_bin.close()
del model_bin del model_bin
del session_bin del session_bin
gc.collect() gc.collect()
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens)
ratio_y=1 ratio_y=1
ratio_x=1 ratio_x=1
@ -1687,11 +1793,15 @@ class Eynollah:
img = resize_image(prediction_bin, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x)) img = resize_image(prediction_bin, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x))
if not self.dir_in:
prediction_regions_org = self.do_prediction(True, img, model_region) prediction_regions_org = self.do_prediction(True, img, model_region)
else:
prediction_regions_org = self.do_prediction(True, img, self.model_region)
prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h ) prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h )
prediction_regions_org=prediction_regions_org[:,:,0] prediction_regions_org=prediction_regions_org[:,:,0]
mask_lines_only=(prediction_regions_org[:,:]==3)*1 mask_lines_only=(prediction_regions_org[:,:]==3)*1
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
@ -1716,7 +1826,7 @@ class Eynollah:
text_regions_p_true=cv2.fillPoly(text_regions_p_true,pts=polygons_of_only_texts, color=(1,1,1)) text_regions_p_true=cv2.fillPoly(text_regions_p_true,pts=polygons_of_only_texts, color=(1,1,1))
if not self.dir_in:
K.clear_session() K.clear_session()
return text_regions_p_true, erosion_hurts, polygons_lines_xml return text_regions_p_true, erosion_hurts, polygons_lines_xml
except: except:
@ -1724,13 +1834,17 @@ class Eynollah:
if self.input_binary: if self.input_binary:
prediction_bin = np.copy(img_org) prediction_bin = np.copy(img_org)
else: else:
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
gc.collect() gc.collect()
if not self.dir_in:
model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization) model_bin, session_bin = self.start_new_session_and_model(self.model_dir_of_binarization)
prediction_bin = self.do_prediction(True, img_org, model_bin) prediction_bin = self.do_prediction(True, img_org, model_bin)
else:
prediction_bin = self.do_prediction(True, img_org, self.model_bin)
prediction_bin = resize_image(prediction_bin, img_height_h, img_width_h ) prediction_bin = resize_image(prediction_bin, img_height_h, img_width_h )
prediction_bin=prediction_bin[:,:,0] prediction_bin=prediction_bin[:,:,0]
@ -1741,26 +1855,29 @@ class Eynollah:
prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2) prediction_bin =np.repeat(prediction_bin[:, :, np.newaxis], 3, axis=2)
if not self.dir_in:
session_bin.close() session_bin.close()
del model_bin del model_bin
del session_bin del session_bin
gc.collect() gc.collect()
if not self.dir_in:
model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens) model_region, session_region = self.start_new_session_and_model(self.model_region_dir_p_ens)
ratio_y=1 ratio_y=1
ratio_x=1 ratio_x=1
img = resize_image(prediction_bin, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x)) img = resize_image(prediction_bin, int(img_org.shape[0]*ratio_y), int(img_org.shape[1]*ratio_x))
if not self.dir_in:
prediction_regions_org = self.do_prediction(True, img, model_region) prediction_regions_org = self.do_prediction(True, img, model_region)
else:
prediction_regions_org = self.do_prediction(True, img, self.model_region)
prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h ) prediction_regions_org = resize_image(prediction_regions_org, img_height_h, img_width_h )
prediction_regions_org=prediction_regions_org[:,:,0] prediction_regions_org=prediction_regions_org[:,:,0]
#mask_lines_only=(prediction_regions_org[:,:]==3)*1 #mask_lines_only=(prediction_regions_org[:,:]==3)*1
if not self.dir_in:
session_region.close() session_region.close()
del model_region del model_region
del session_region del session_region
@ -1807,6 +1924,7 @@ class Eynollah:
text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1)) text_regions_p_true = cv2.fillPoly(text_regions_p_true, pts = polygons_of_only_texts, color=(1,1,1))
erosion_hurts = True erosion_hurts = True
if not self.dir_in:
K.clear_session() K.clear_session()
return text_regions_p_true, erosion_hurts, polygons_lines_xml return text_regions_p_true, erosion_hurts, polygons_lines_xml
@ -2409,7 +2527,7 @@ class Eynollah:
##print(img_only_regions.shape,'img_only_regions') ##print(img_only_regions.shape,'img_only_regions')
##plt.imshow(img_only_regions[:,:]) ##plt.imshow(img_only_regions[:,:])
##plt.show() ##plt.show()
num_col, _ = find_num_col(img_only_regions, num_col_classifier, self.tables, multiplier=6.0) ##num_col, _ = find_num_col(img_only_regions, num_col_classifier, self.tables, multiplier=6.0)
try: try:
num_col, _ = find_num_col(img_only_regions, num_col_classifier, self.tables, multiplier=6.0) num_col, _ = find_num_col(img_only_regions, num_col_classifier, self.tables, multiplier=6.0)
num_col = num_col + 1 num_col = num_col + 1
@ -2468,6 +2586,7 @@ class Eynollah:
self.logger.info("Resizing and enhancing image...") self.logger.info("Resizing and enhancing image...")
is_image_enhanced, img_org, img_res, num_col_classifier, num_column_is_classified, img_bin = self.resize_and_enhance_image_with_column_classifier(light_version) is_image_enhanced, img_org, img_res, num_col_classifier, num_column_is_classified, img_bin = self.resize_and_enhance_image_with_column_classifier(light_version)
self.logger.info("Image was %senhanced.", '' if is_image_enhanced else 'not ') self.logger.info("Image was %senhanced.", '' if is_image_enhanced else 'not ')
if not self.dir_in:
K.clear_session() K.clear_session()
scale = 1 scale = 1
if is_image_enhanced: if is_image_enhanced:
@ -2492,6 +2611,7 @@ class Eynollah:
scaler_h_textline = 1 # 1.2#1.2 scaler_h_textline = 1 # 1.2#1.2
scaler_w_textline = 1 # 0.9#1 scaler_w_textline = 1 # 0.9#1
textline_mask_tot_ea, _ = self.textline_contours(image_page, True, scaler_h_textline, scaler_w_textline) textline_mask_tot_ea, _ = self.textline_contours(image_page, True, scaler_h_textline, scaler_w_textline)
if not self.dir_in:
K.clear_session() K.clear_session()
if self.plotter: if self.plotter:
self.plotter.save_plot_of_textlines(textline_mask_tot_ea, image_page) self.plotter.save_plot_of_textlines(textline_mask_tot_ea, image_page)
@ -2554,6 +2674,7 @@ class Eynollah:
if np.abs(slope_deskew) >= SLOPE_THRESHOLD: if np.abs(slope_deskew) >= SLOPE_THRESHOLD:
_, _, matrix_of_lines_ch_d, splitter_y_new_d, _ = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2), num_col_classifier, self.tables, pixel_lines) _, _, matrix_of_lines_ch_d, splitter_y_new_d, _ = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2), num_col_classifier, self.tables, pixel_lines)
if not self.dir_in:
K.clear_session() K.clear_session()
self.logger.info("num_col_classifier: %s", num_col_classifier) self.logger.info("num_col_classifier: %s", num_col_classifier)
@ -2619,7 +2740,7 @@ class Eynollah:
pixel_img = 10 pixel_img = 10
contours_tables = return_contours_of_interested_region(text_regions_p, pixel_img, min_area_mar) contours_tables = return_contours_of_interested_region(text_regions_p, pixel_img, min_area_mar)
if not self.dir_in:
K.clear_session() K.clear_session()
self.logger.debug('exit run_boxes_no_full_layout') self.logger.debug('exit run_boxes_no_full_layout')
return polygons_of_images, img_revised_tab, text_regions_p_1_n, textline_mask_tot_d, regions_without_separators_d, boxes, boxes_d, polygons_of_marginals, contours_tables return polygons_of_images, img_revised_tab, text_regions_p_1_n, textline_mask_tot_d, regions_without_separators_d, boxes, boxes_d, polygons_of_marginals, contours_tables
@ -2651,6 +2772,7 @@ class Eynollah:
if np.abs(slope_deskew) >= SLOPE_THRESHOLD: if np.abs(slope_deskew) >= SLOPE_THRESHOLD:
num_col_d, peaks_neg_fin_d, matrix_of_lines_ch_d, splitter_y_new_d, seperators_closeup_n_d = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2),num_col_classifier, self.tables, pixel_lines) num_col_d, peaks_neg_fin_d, matrix_of_lines_ch_d, splitter_y_new_d, seperators_closeup_n_d = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2),num_col_classifier, self.tables, pixel_lines)
if not self.dir_in:
K.clear_session() K.clear_session()
gc.collect() gc.collect()
@ -2718,7 +2840,7 @@ class Eynollah:
text_regions_p[:, :][text_regions_p[:, :] == 2] = 5 text_regions_p[:, :][text_regions_p[:, :] == 2] = 5
text_regions_p[:, :][text_regions_p[:, :] == 3] = 6 text_regions_p[:, :][text_regions_p[:, :] == 3] = 6
text_regions_p[:, :][text_regions_p[:, :] == 4] = 8 text_regions_p[:, :][text_regions_p[:, :] == 4] = 8
if not self.dir_in:
K.clear_session() K.clear_session()
image_page = image_page.astype(np.uint8) image_page = image_page.astype(np.uint8)
@ -2726,6 +2848,7 @@ class Eynollah:
text_regions_p[:,:][regions_fully[:,:,0]==6]=6 text_regions_p[:,:][regions_fully[:,:,0]==6]=6
regions_fully_only_drop = put_drop_out_from_only_drop_model(regions_fully_only_drop, text_regions_p) regions_fully_only_drop = put_drop_out_from_only_drop_model(regions_fully_only_drop, text_regions_p)
regions_fully[:, :, 0][regions_fully_only_drop[:, :, 0] == 4] = 4 regions_fully[:, :, 0][regions_fully_only_drop[:, :, 0] == 4] = 4
if not self.dir_in:
K.clear_session() K.clear_session()
# plt.imshow(regions_fully[:,:,0]) # plt.imshow(regions_fully[:,:,0])
@ -2733,6 +2856,7 @@ class Eynollah:
regions_fully = putt_bb_of_drop_capitals_of_model_in_patches_in_layout(regions_fully) regions_fully = putt_bb_of_drop_capitals_of_model_in_patches_in_layout(regions_fully)
# plt.imshow(regions_fully[:,:,0]) # plt.imshow(regions_fully[:,:,0])
# plt.show() # plt.show()
if not self.dir_in:
K.clear_session() K.clear_session()
regions_fully_np, _ = self.extract_text_regions(image_page, False, cols=num_col_classifier) regions_fully_np, _ = self.extract_text_regions(image_page, False, cols=num_col_classifier)
# plt.imshow(regions_fully_np[:,:,0]) # plt.imshow(regions_fully_np[:,:,0])
@ -2744,6 +2868,7 @@ class Eynollah:
# plt.imshow(regions_fully_np[:,:,0]) # plt.imshow(regions_fully_np[:,:,0])
# plt.show() # plt.show()
if not self.dir_in:
K.clear_session() K.clear_session()
# plt.imshow(regions_fully[:,:,0]) # plt.imshow(regions_fully[:,:,0])
# plt.show() # plt.show()
@ -2769,20 +2894,35 @@ class Eynollah:
regions_without_separators_d = None regions_without_separators_d = None
if not self.tables: if not self.tables:
regions_without_separators = (text_regions_p[:, :] == 1) * 1 regions_without_separators = (text_regions_p[:, :] == 1) * 1
if not self.dir_in:
K.clear_session() K.clear_session()
img_revised_tab = np.copy(text_regions_p[:, :]) img_revised_tab = np.copy(text_regions_p[:, :])
polygons_of_images = return_contours_of_interested_region(img_revised_tab, 5) polygons_of_images = return_contours_of_interested_region(img_revised_tab, 5)
self.logger.debug('exit run_boxes_full_layout') self.logger.debug('exit run_boxes_full_layout')
return polygons_of_images, img_revised_tab, text_regions_p_1_n, textline_mask_tot_d, regions_without_separators_d, regions_fully, regions_without_separators, polygons_of_marginals, contours_tables return polygons_of_images, img_revised_tab, text_regions_p_1_n, textline_mask_tot_d, regions_without_separators_d, regions_fully, regions_without_separators, polygons_of_marginals, contours_tables
def our_load_model(self, model_file):
model = load_model(model_file, compile=False)
return model
def run(self): def run(self):
""" """
Get image and scales, then extract the page of scanned image Get image and scales, then extract the page of scanned image
""" """
self.logger.debug("enter run") self.logger.debug("enter run")
t0_tot = time.time()
if not self.dir_in:
self.ls_imgs = [1]
for img_name in self.ls_imgs:
t0 = time.time() t0 = time.time()
if self.dir_in:
self.reset_file_name_dir(os.path.join(self.dir_in,img_name))
img_res, is_image_enhanced, num_col_classifier, num_column_is_classified = self.run_enhancement(self.light_version) img_res, is_image_enhanced, num_col_classifier, num_column_is_classified = self.run_enhancement(self.light_version)
self.logger.info("Enhancing took %.1fs ", time.time() - t0) self.logger.info("Enhancing took %.1fs ", time.time() - t0)
@ -2790,11 +2930,11 @@ class Eynollah:
t1 = time.time() t1 = time.time()
if self.light_version: if self.light_version:
text_regions_p_1 ,erosion_hurts, polygons_lines_xml, textline_mask_tot_ea = self.get_regions_from_xy_2models_light(img_res, is_image_enhanced, num_col_classifier) text_regions_p_1 ,erosion_hurts, polygons_lines_xml, textline_mask_tot_ea = self.get_regions_from_xy_2models_light(img_res, is_image_enhanced, num_col_classifier)
slope_deskew, slope_first = self.run_deskew(textline_mask_tot_ea) slope_deskew, slope_first = self.run_deskew(textline_mask_tot_ea)
#self.logger.info("Textregion detection took %.1fs ", time.time() - t1t)
num_col, num_col_classifier, img_only_regions, page_coord, image_page, mask_images, mask_lines, text_regions_p_1, cont_page, table_prediction, textline_mask_tot_ea = \ num_col, num_col_classifier, img_only_regions, page_coord, image_page, mask_images, mask_lines, text_regions_p_1, cont_page, table_prediction, textline_mask_tot_ea = \
self.run_graphics_and_columns_light(text_regions_p_1, textline_mask_tot_ea, num_col_classifier, num_column_is_classified, erosion_hurts) self.run_graphics_and_columns_light(text_regions_p_1, textline_mask_tot_ea, num_col_classifier, num_column_is_classified, erosion_hurts)
#self.logger.info("run graphics %.1fs ", time.time() - t1t)
else: else:
text_regions_p_1 ,erosion_hurts, polygons_lines_xml = self.get_regions_from_xy_2models(img_res, is_image_enhanced, num_col_classifier) text_regions_p_1 ,erosion_hurts, polygons_lines_xml = self.get_regions_from_xy_2models(img_res, is_image_enhanced, num_col_classifier)
self.logger.info("Textregion detection took %.1fs ", time.time() - t1) self.logger.info("Textregion detection took %.1fs ", time.time() - t1)
@ -2803,7 +2943,7 @@ class Eynollah:
num_col, num_col_classifier, img_only_regions, page_coord, image_page, mask_images, mask_lines, text_regions_p_1, cont_page, table_prediction = \ num_col, num_col_classifier, img_only_regions, page_coord, image_page, mask_images, mask_lines, text_regions_p_1, cont_page, table_prediction = \
self.run_graphics_and_columns(text_regions_p_1, num_col_classifier, num_column_is_classified, erosion_hurts) self.run_graphics_and_columns(text_regions_p_1, num_col_classifier, num_column_is_classified, erosion_hurts)
self.logger.info("Graphics detection took %.1fs ", time.time() - t1) self.logger.info("Graphics detection took %.1fs ", time.time() - t1)
self.logger.info('cont_page %s', cont_page) #self.logger.info('cont_page %s', cont_page)
if not num_col: if not num_col:
self.logger.info("No columns detected, outputting an empty PAGE-XML") self.logger.info("No columns detected, outputting an empty PAGE-XML")
@ -2835,6 +2975,7 @@ class Eynollah:
if np.abs(slope_deskew) >= SLOPE_THRESHOLD: if np.abs(slope_deskew) >= SLOPE_THRESHOLD:
text_only_d = ((text_regions_p_1_n[:, :] == 1)) * 1 text_only_d = ((text_regions_p_1_n[:, :] == 1)) * 1
min_con_area = 0.000005 min_con_area = 0.000005
if np.abs(slope_deskew) >= SLOPE_THRESHOLD: if np.abs(slope_deskew) >= SLOPE_THRESHOLD:
contours_only_text, hir_on_text = return_contours_of_image(text_only) contours_only_text, hir_on_text = return_contours_of_image(text_only)
@ -2843,7 +2984,7 @@ class Eynollah:
if len(contours_only_text_parent) > 0: if len(contours_only_text_parent) > 0:
areas_cnt_text = np.array([cv2.contourArea(contours_only_text_parent[j]) for j in range(len(contours_only_text_parent))]) areas_cnt_text = np.array([cv2.contourArea(contours_only_text_parent[j]) for j in range(len(contours_only_text_parent))])
areas_cnt_text = areas_cnt_text / float(text_only.shape[0] * text_only.shape[1]) areas_cnt_text = areas_cnt_text / float(text_only.shape[0] * text_only.shape[1])
self.logger.info('areas_cnt_text %s', areas_cnt_text) #self.logger.info('areas_cnt_text %s', areas_cnt_text)
contours_biggest = contours_only_text_parent[np.argmax(areas_cnt_text)] contours_biggest = contours_only_text_parent[np.argmax(areas_cnt_text)]
contours_only_text_parent = [contours_only_text_parent[jz] for jz in range(len(contours_only_text_parent)) if areas_cnt_text[jz] > min_con_area] contours_only_text_parent = [contours_only_text_parent[jz] for jz in range(len(contours_only_text_parent)) if areas_cnt_text[jz] > min_con_area]
areas_cnt_text_parent = [areas_cnt_text[jz] for jz in range(len(areas_cnt_text)) if areas_cnt_text[jz] > min_con_area] areas_cnt_text_parent = [areas_cnt_text[jz] for jz in range(len(areas_cnt_text)) if areas_cnt_text[jz] > min_con_area]
@ -2914,7 +3055,6 @@ class Eynollah:
contours_only_text_parent_d_ordered = [] contours_only_text_parent_d_ordered = []
contours_only_text_parent_d = [] contours_only_text_parent_d = []
contours_only_text_parent = [] contours_only_text_parent = []
else: else:
contours_only_text, hir_on_text = return_contours_of_image(text_only) contours_only_text, hir_on_text = return_contours_of_image(text_only)
contours_only_text_parent = return_parent_contours(contours_only_text, hir_on_text) contours_only_text_parent = return_parent_contours(contours_only_text, hir_on_text)
@ -2933,11 +3073,14 @@ class Eynollah:
cx_bigest_big, cy_biggest_big, _, _, _, _, _ = find_new_features_of_contours([contours_biggest]) cx_bigest_big, cy_biggest_big, _, _, _, _, _ = find_new_features_of_contours([contours_biggest])
cx_bigest, cy_biggest, _, _, _, _, _ = find_new_features_of_contours(contours_only_text_parent) cx_bigest, cy_biggest, _, _, _, _, _ = find_new_features_of_contours(contours_only_text_parent)
self.logger.debug('areas_cnt_text_parent %s', areas_cnt_text_parent) #self.logger.debug('areas_cnt_text_parent %s', areas_cnt_text_parent)
# self.logger.debug('areas_cnt_text_parent_d %s', areas_cnt_text_parent_d) # self.logger.debug('areas_cnt_text_parent_d %s', areas_cnt_text_parent_d)
# self.logger.debug('len(contours_only_text_parent) %s', len(contours_only_text_parent_d)) # self.logger.debug('len(contours_only_text_parent) %s', len(contours_only_text_parent_d))
else: else:
pass pass
if self.light_version:
txt_con_org = get_textregion_contours_in_org_image_light(contours_only_text_parent, self.image, slope_first)
else:
txt_con_org = get_textregion_contours_in_org_image(contours_only_text_parent, self.image, slope_first) txt_con_org = get_textregion_contours_in_org_image(contours_only_text_parent, self.image, slope_first)
boxes_text, _ = get_text_region_boxes_by_given_contours(contours_only_text_parent) boxes_text, _ = get_text_region_boxes_by_given_contours(contours_only_text_parent)
boxes_marginals, _ = get_text_region_boxes_by_given_contours(polygons_of_marginals) boxes_marginals, _ = get_text_region_boxes_by_given_contours(polygons_of_marginals)
@ -2956,26 +3099,33 @@ class Eynollah:
all_found_texline_polygons = small_textlines_to_parent_adherence2(all_found_texline_polygons, textline_mask_tot_ea, num_col_classifier) all_found_texline_polygons = small_textlines_to_parent_adherence2(all_found_texline_polygons, textline_mask_tot_ea, num_col_classifier)
all_found_texline_polygons_marginals, boxes_marginals, _, polygons_of_marginals, all_box_coord_marginals, _, slopes_marginals = self.get_slopes_and_deskew_new_curved(polygons_of_marginals, polygons_of_marginals, cv2.erode(textline_mask_tot_ea, kernel=KERNEL, iterations=1), image_page_rotated, boxes_marginals, text_only, num_col_classifier, scale_param, slope_deskew) all_found_texline_polygons_marginals, boxes_marginals, _, polygons_of_marginals, all_box_coord_marginals, _, slopes_marginals = self.get_slopes_and_deskew_new_curved(polygons_of_marginals, polygons_of_marginals, cv2.erode(textline_mask_tot_ea, kernel=KERNEL, iterations=1), image_page_rotated, boxes_marginals, text_only, num_col_classifier, scale_param, slope_deskew)
all_found_texline_polygons_marginals = small_textlines_to_parent_adherence2(all_found_texline_polygons_marginals, textline_mask_tot_ea, num_col_classifier) all_found_texline_polygons_marginals = small_textlines_to_parent_adherence2(all_found_texline_polygons_marginals, textline_mask_tot_ea, num_col_classifier)
if not self.dir_in:
K.clear_session() K.clear_session()
if self.full_layout: if self.full_layout:
if np.abs(slope_deskew) >= SLOPE_THRESHOLD: if np.abs(slope_deskew) >= SLOPE_THRESHOLD:
contours_only_text_parent_d_ordered = list(np.array(contours_only_text_parent_d_ordered)[index_by_text_par_con]) contours_only_text_parent_d_ordered = list(np.array(contours_only_text_parent_d_ordered)[index_by_text_par_con])
if self.light_version:
text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header_light(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered)
else:
text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered) text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered)
else: else:
#takes long timee
contours_only_text_parent_d_ordered = None contours_only_text_parent_d_ordered = None
if self.light_version:
text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header_light(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered)
else:
text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered) text_regions_p, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, slopes, slopes_h, contours_only_text_parent_d_ordered, contours_only_text_parent_h_d_ordered = check_any_text_region_in_model_one_is_main_or_header(text_regions_p, regions_fully, contours_only_text_parent, all_box_coord, all_found_texline_polygons, slopes, contours_only_text_parent_d_ordered)
if self.plotter: if self.plotter:
self.plotter.save_plot_of_layout(text_regions_p, image_page) self.plotter.save_plot_of_layout(text_regions_p, image_page)
self.plotter.save_plot_of_layout_all(text_regions_p, image_page) self.plotter.save_plot_of_layout_all(text_regions_p, image_page)
if not self.dir_in:
K.clear_session() K.clear_session()
pixel_img = 4 pixel_img = 4
polygons_of_drop_capitals = return_contours_of_interested_region_by_min_size(text_regions_p, pixel_img) polygons_of_drop_capitals = return_contours_of_interested_region_by_min_size(text_regions_p, pixel_img)
all_found_texline_polygons = adhere_drop_capital_region_into_corresponding_textline(text_regions_p, polygons_of_drop_capitals, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, kernel=KERNEL, curved_line=self.curved_line) all_found_texline_polygons = adhere_drop_capital_region_into_corresponding_textline(text_regions_p, polygons_of_drop_capitals, contours_only_text_parent, contours_only_text_parent_h, all_box_coord, all_box_coord_h, all_found_texline_polygons, all_found_texline_polygons_h, kernel=KERNEL, curved_line=self.curved_line)
# print(len(contours_only_text_parent_h),len(contours_only_text_parent_h_d_ordered),'contours_only_text_parent_h')
pixel_lines = 6 pixel_lines = 6
@ -2990,45 +3140,15 @@ class Eynollah:
else: else:
_, _, matrix_of_lines_ch_d, splitter_y_new_d, _ = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2), num_col_classifier, self.tables, pixel_lines) _, _, matrix_of_lines_ch_d, splitter_y_new_d, _ = find_number_of_columns_in_document(np.repeat(text_regions_p_1_n[:, :, np.newaxis], 3, axis=2), num_col_classifier, self.tables, pixel_lines)
# print(peaks_neg_fin,peaks_neg_fin_d,'num_col2')
# print(splitter_y_new,splitter_y_new_d,'num_col_classifier')
# print(matrix_of_lines_ch.shape,matrix_of_lines_ch_d.shape,'matrix_of_lines_ch')
if num_col_classifier >= 3: if num_col_classifier >= 3:
if np.abs(slope_deskew) < SLOPE_THRESHOLD: if np.abs(slope_deskew) < SLOPE_THRESHOLD:
regions_without_separators = regions_without_separators.astype(np.uint8) regions_without_separators = regions_without_separators.astype(np.uint8)
regions_without_separators = cv2.erode(regions_without_separators[:, :], KERNEL, iterations=6) regions_without_separators = cv2.erode(regions_without_separators[:, :], KERNEL, iterations=6)
#regions_without_separators_0 = regions_without_separators[:, :].sum(axis=0)
#meda_n_updown = regions_without_separators_0[len(regions_without_separators_0) :: -1]
#first_nonzero = next((i for i, x in enumerate(regions_without_separators_0) if x), 0)
#last_nonzero = next((i for i, x in enumerate(meda_n_updown) if x), 0)
#last_nonzero = len(regions_without_separators_0) - last_nonzero
#random_pixels_for_image = np.random.randn(regions_without_separators.shape[0], regions_without_separators.shape[1])
#random_pixels_for_image[random_pixels_for_image < -0.5] = 0
#random_pixels_for_image[random_pixels_for_image != 0] = 1
#regions_without_separators[(random_pixels_for_image[:, :] == 1) & (text_regions_p[:, :] == 5)] = 1
#regions_without_separators[:, 0:first_nonzero] = 0
#regions_without_separators[:, last_nonzero:] = 0
else: else:
regions_without_separators_d = regions_without_separators_d.astype(np.uint8) regions_without_separators_d = regions_without_separators_d.astype(np.uint8)
regions_without_separators_d = cv2.erode(regions_without_separators_d[:, :], KERNEL, iterations=6) regions_without_separators_d = cv2.erode(regions_without_separators_d[:, :], KERNEL, iterations=6)
#regions_without_separators_0 = regions_without_separators_d[:, :].sum(axis=0)
#meda_n_updown = regions_without_separators_0[len(regions_without_separators_0) :: -1]
#first_nonzero = next((i for i, x in enumerate(regions_without_separators_0) if x), 0)
#last_nonzero = next((i for i, x in enumerate(meda_n_updown) if x), 0)
#last_nonzero = len(regions_without_separators_0) - last_nonzero
#random_pixels_for_image = np.random.randn(regions_without_separators_d.shape[0], regions_without_separators_d.shape[1])
#random_pixels_for_image[random_pixels_for_image < -0.5] = 0
#random_pixels_for_image[random_pixels_for_image != 0] = 1
##regions_without_separators_d[(random_pixels_for_image[:, :] == 1) & (text_regions_p_1_n[:, :] == 5)] = 1
#regions_without_separators_d[:, 0:first_nonzero] = 0
#regions_without_separators_d[:, last_nonzero:] = 0
if np.abs(slope_deskew) < SLOPE_THRESHOLD: if np.abs(slope_deskew) < SLOPE_THRESHOLD:
boxes, peaks_neg_tot_tables = return_boxes_of_images_by_order_of_reading_new(splitter_y_new, regions_without_separators, matrix_of_lines_ch, num_col_classifier, erosion_hurts, self.tables) boxes, peaks_neg_tot_tables = return_boxes_of_images_by_order_of_reading_new(splitter_y_new, regions_without_separators, matrix_of_lines_ch, num_col_classifier, erosion_hurts, self.tables)
@ -3037,7 +3157,7 @@ class Eynollah:
if self.plotter: if self.plotter:
self.plotter.write_images_into_directory(polygons_of_images, image_page) self.plotter.write_images_into_directory(polygons_of_images, image_page)
t_order = time.time()
if self.full_layout: if self.full_layout:
if np.abs(slope_deskew) < SLOPE_THRESHOLD: if np.abs(slope_deskew) < SLOPE_THRESHOLD:
order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent, contours_only_text_parent_h, boxes, textline_mask_tot) order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent, contours_only_text_parent_h, boxes, textline_mask_tot)
@ -3046,7 +3166,7 @@ class Eynollah:
pcgts = self.writer.build_pagexml_full_layout(contours_only_text_parent, contours_only_text_parent_h, page_coord, order_text_new, id_of_texts_tot, all_found_texline_polygons, all_found_texline_polygons_h, all_box_coord, all_box_coord_h, polygons_of_images, contours_tables, polygons_of_drop_capitals, polygons_of_marginals, all_found_texline_polygons_marginals, all_box_coord_marginals, slopes, slopes_h, slopes_marginals, cont_page, polygons_lines_xml) pcgts = self.writer.build_pagexml_full_layout(contours_only_text_parent, contours_only_text_parent_h, page_coord, order_text_new, id_of_texts_tot, all_found_texline_polygons, all_found_texline_polygons_h, all_box_coord, all_box_coord_h, polygons_of_images, contours_tables, polygons_of_drop_capitals, polygons_of_marginals, all_found_texline_polygons_marginals, all_box_coord_marginals, slopes, slopes_h, slopes_marginals, cont_page, polygons_lines_xml)
self.logger.info("Job done in %.1fs", time.time() - t0) self.logger.info("Job done in %.1fs", time.time() - t0)
return pcgts ##return pcgts
else: else:
contours_only_text_parent_h = None contours_only_text_parent_h = None
if np.abs(slope_deskew) < SLOPE_THRESHOLD: if np.abs(slope_deskew) < SLOPE_THRESHOLD:
@ -3056,4 +3176,7 @@ class Eynollah:
order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent_d_ordered, contours_only_text_parent_h, boxes_d, textline_mask_tot_d) order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent_d_ordered, contours_only_text_parent_h, boxes_d, textline_mask_tot_d)
pcgts = self.writer.build_pagexml_no_full_layout(txt_con_org, page_coord, order_text_new, id_of_texts_tot, all_found_texline_polygons, all_box_coord, polygons_of_images, polygons_of_marginals, all_found_texline_polygons_marginals, all_box_coord_marginals, slopes, slopes_marginals, cont_page, polygons_lines_xml, contours_tables) pcgts = self.writer.build_pagexml_no_full_layout(txt_con_org, page_coord, order_text_new, id_of_texts_tot, all_found_texline_polygons, all_box_coord, polygons_of_images, polygons_of_marginals, all_found_texline_polygons_marginals, all_box_coord_marginals, slopes, slopes_marginals, cont_page, polygons_lines_xml, contours_tables)
self.logger.info("Job done in %.1fs", time.time() - t0) self.logger.info("Job done in %.1fs", time.time() - t0)
return pcgts ##return pcgts
self.writer.write_pagexml(pcgts)
#self.logger.info("Job done in %.1fs", time.time() - t0)
self.logger.info("All jobs done in %.1fs", time.time() - t0_tot)

80
qurator/eynollah/utils/__init__.py
View file

@ -797,6 +797,76 @@ def putt_bb_of_drop_capitals_of_model_in_patches_in_layout(layout_in_patch):
return layout_in_patch return layout_in_patch
def check_any_text_region_in_model_one_is_main_or_header(regions_model_1,regions_model_full,contours_only_text_parent,all_box_coord,all_found_texline_polygons,slopes,contours_only_text_parent_d_ordered): def check_any_text_region_in_model_one_is_main_or_header(regions_model_1,regions_model_full,contours_only_text_parent,all_box_coord,all_found_texline_polygons,slopes,contours_only_text_parent_d_ordered):
cx_main,cy_main ,x_min_main , x_max_main, y_min_main ,y_max_main,y_corr_x_min_from_argmin=find_new_features_of_contours(contours_only_text_parent)
length_con=x_max_main-x_min_main
height_con=y_max_main-y_min_main
all_found_texline_polygons_main=[]
all_found_texline_polygons_head=[]
all_box_coord_main=[]
all_box_coord_head=[]
slopes_main=[]
slopes_head=[]
contours_only_text_parent_main=[]
contours_only_text_parent_head=[]
contours_only_text_parent_main_d=[]
contours_only_text_parent_head_d=[]
for ii in range(len(contours_only_text_parent)):
con=contours_only_text_parent[ii]
img=np.zeros((regions_model_1.shape[0],regions_model_1.shape[1],3))
img = cv2.fillPoly(img, pts=[con], color=(255, 255, 255))
all_pixels=((img[:,:,0]==255)*1).sum()
pixels_header=( ( (img[:,:,0]==255) & (regions_model_full[:,:,0]==2) )*1 ).sum()
pixels_main=all_pixels-pixels_header
if (pixels_header>=pixels_main) and ( (length_con[ii]/float(height_con[ii]) )>=1.3 ):
regions_model_1[:,:][(regions_model_1[:,:]==1) & (img[:,:,0]==255) ]=2
contours_only_text_parent_head.append(con)
if contours_only_text_parent_d_ordered is not None:
contours_only_text_parent_head_d.append(contours_only_text_parent_d_ordered[ii])
all_box_coord_head.append(all_box_coord[ii])
slopes_head.append(slopes[ii])
all_found_texline_polygons_head.append(all_found_texline_polygons[ii])
else:
regions_model_1[:,:][(regions_model_1[:,:]==1) & (img[:,:,0]==255) ]=1
contours_only_text_parent_main.append(con)
if contours_only_text_parent_d_ordered is not None:
contours_only_text_parent_main_d.append(contours_only_text_parent_d_ordered[ii])
all_box_coord_main.append(all_box_coord[ii])
slopes_main.append(slopes[ii])
all_found_texline_polygons_main.append(all_found_texline_polygons[ii])
#print(all_pixels,pixels_main,pixels_header)
return regions_model_1,contours_only_text_parent_main,contours_only_text_parent_head,all_box_coord_main,all_box_coord_head,all_found_texline_polygons_main,all_found_texline_polygons_head,slopes_main,slopes_head,contours_only_text_parent_main_d,contours_only_text_parent_head_d
def check_any_text_region_in_model_one_is_main_or_header_light(regions_model_1,regions_model_full,contours_only_text_parent,all_box_coord,all_found_texline_polygons,slopes,contours_only_text_parent_d_ordered):
### to make it faster
h_o = regions_model_1.shape[0]
w_o = regions_model_1.shape[1]
regions_model_1 = cv2.resize(regions_model_1, (int(regions_model_1.shape[1]/3.), int(regions_model_1.shape[0]/3.)), interpolation=cv2.INTER_NEAREST)
regions_model_full = cv2.resize(regions_model_full, (int(regions_model_full.shape[1]/3.), int(regions_model_full.shape[0]/3.)), interpolation=cv2.INTER_NEAREST)
contours_only_text_parent = [ (i/3.).astype(np.int32) for i in contours_only_text_parent]
###
cx_main,cy_main ,x_min_main , x_max_main, y_min_main ,y_max_main,y_corr_x_min_from_argmin=find_new_features_of_contours(contours_only_text_parent) cx_main,cy_main ,x_min_main , x_max_main, y_min_main ,y_max_main,y_corr_x_min_from_argmin=find_new_features_of_contours(contours_only_text_parent)
length_con=x_max_main-x_min_main length_con=x_max_main-x_min_main
@ -853,8 +923,14 @@ def check_any_text_region_in_model_one_is_main_or_header(regions_model_1,regions
#plt.imshow(img[:,:,0]) ### to make it faster
#plt.show()
regions_model_1 = cv2.resize(regions_model_1, (w_o, h_o), interpolation=cv2.INTER_NEAREST)
#regions_model_full = cv2.resize(img, (int(regions_model_full.shape[1]/3.), int(regions_model_full.shape[0]/3.)), interpolation=cv2.INTER_NEAREST)
contours_only_text_parent_head = [ (i*3.).astype(np.int32) for i in contours_only_text_parent_head]
contours_only_text_parent_main = [ (i*3.).astype(np.int32) for i in contours_only_text_parent_main]
###
return regions_model_1,contours_only_text_parent_main,contours_only_text_parent_head,all_box_coord_main,all_box_coord_head,all_found_texline_polygons_main,all_found_texline_polygons_head,slopes_main,slopes_head,contours_only_text_parent_main_d,contours_only_text_parent_head_d return regions_model_1,contours_only_text_parent_main,contours_only_text_parent_head,all_box_coord_main,all_box_coord_head,all_found_texline_polygons_main,all_found_texline_polygons_head,slopes_main,slopes_head,contours_only_text_parent_main_d,contours_only_text_parent_head_d
def small_textlines_to_parent_adherence2(textlines_con, textline_iamge, num_col): def small_textlines_to_parent_adherence2(textlines_con, textline_iamge, num_col):

133
qurator/eynollah/utils/contour.py
View file

@ -3,7 +3,8 @@ import numpy as np
from shapely import geometry from shapely import geometry
from .rotate import rotate_image, rotation_image_new from .rotate import rotate_image, rotation_image_new
from multiprocessing import Process, Queue, cpu_count
from multiprocessing import Pool
def contours_in_same_horizon(cy_main_hor): def contours_in_same_horizon(cy_main_hor):
X1 = np.zeros((len(cy_main_hor), len(cy_main_hor))) X1 = np.zeros((len(cy_main_hor), len(cy_main_hor)))
X2 = np.zeros((len(cy_main_hor), len(cy_main_hor))) X2 = np.zeros((len(cy_main_hor), len(cy_main_hor)))
@ -147,6 +148,96 @@ def return_contours_of_interested_region(region_pre_p, pixel, min_area=0.0002):
return contours_imgs return contours_imgs
def do_work_of_contours_in_image(queue_of_all_params, contours_per_process, indexes_r_con_per_pro, img, slope_first):
cnts_org_per_each_subprocess = []
index_by_text_region_contours = []
for mv in range(len(contours_per_process)):
index_by_text_region_contours.append(indexes_r_con_per_pro[mv])
img_copy = np.zeros(img.shape)
img_copy = cv2.fillPoly(img_copy, pts=[contours_per_process[mv]], color=(1, 1, 1))
img_copy = rotation_image_new(img_copy, -slope_first)
img_copy = img_copy.astype(np.uint8)
imgray = cv2.cvtColor(img_copy, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(imgray, 0, 255, 0)
cont_int, _ = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
cont_int[0][:, 0, 0] = cont_int[0][:, 0, 0] + np.abs(img_copy.shape[1] - img.shape[1])
cont_int[0][:, 0, 1] = cont_int[0][:, 0, 1] + np.abs(img_copy.shape[0] - img.shape[0])
cnts_org_per_each_subprocess.append(cont_int[0])
queue_of_all_params.put([ cnts_org_per_each_subprocess, index_by_text_region_contours])
def get_textregion_contours_in_org_image_multi(cnts, img, slope_first):
num_cores = cpu_count()
queue_of_all_params = Queue()
processes = []
nh = np.linspace(0, len(cnts), num_cores + 1)
indexes_by_text_con = np.array(range(len(cnts)))
for i in range(num_cores):
contours_per_process = cnts[int(nh[i]) : int(nh[i + 1])]
indexes_text_con_per_process = indexes_by_text_con[int(nh[i]) : int(nh[i + 1])]
processes.append(Process(target=do_work_of_contours_in_image, args=(queue_of_all_params, contours_per_process, indexes_text_con_per_process, img,slope_first )))
for i in range(num_cores):
processes[i].start()
cnts_org = []
all_index_text_con = []
for i in range(num_cores):
list_all_par = queue_of_all_params.get(True)
contours_for_sub_process = list_all_par[0]
indexes_for_sub_process = list_all_par[1]
for j in range(len(contours_for_sub_process)):
cnts_org.append(contours_for_sub_process[j])
all_index_text_con.append(indexes_for_sub_process[j])
for i in range(num_cores):
processes[i].join()
print(all_index_text_con)
return cnts_org
def loop_contour_image(index_l, cnts,img, slope_first):
img_copy = np.zeros(img.shape)
img_copy = cv2.fillPoly(img_copy, pts=[cnts[index_l]], color=(1, 1, 1))
# plt.imshow(img_copy)
# plt.show()
# print(img.shape,'img')
img_copy = rotation_image_new(img_copy, -slope_first)
##print(img_copy.shape,'img_copy')
# plt.imshow(img_copy)
# plt.show()
img_copy = img_copy.astype(np.uint8)
imgray = cv2.cvtColor(img_copy, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(imgray, 0, 255, 0)
cont_int, _ = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
cont_int[0][:, 0, 0] = cont_int[0][:, 0, 0] + np.abs(img_copy.shape[1] - img.shape[1])
cont_int[0][:, 0, 1] = cont_int[0][:, 0, 1] + np.abs(img_copy.shape[0] - img.shape[0])
# print(np.shape(cont_int[0]))
return cont_int[0]
def get_textregion_contours_in_org_image_multi2(cnts, img, slope_first):
cnts_org = []
# print(cnts,'cnts')
with Pool(cpu_count()) as p:
cnts_org = p.starmap(loop_contour_image, [(index_l,cnts, img,slope_first) for index_l in range(len(cnts))])
print(len(cnts_org),'lendiha')
return cnts_org
def get_textregion_contours_in_org_image(cnts, img, slope_first): def get_textregion_contours_in_org_image(cnts, img, slope_first):
cnts_org = [] cnts_org = []
@ -175,11 +266,43 @@ def get_textregion_contours_in_org_image(cnts, img, slope_first):
# print(np.shape(cont_int[0])) # print(np.shape(cont_int[0]))
cnts_org.append(cont_int[0]) cnts_org.append(cont_int[0])
# print(cnts_org,'cnts_org') return cnts_org
def get_textregion_contours_in_org_image_light(cnts, img, slope_first):
h_o = img.shape[0]
w_o = img.shape[1]
img = cv2.resize(img, (int(img.shape[1]/3.), int(img.shape[0]/3.)), interpolation=cv2.INTER_NEAREST)
##cnts = list( (np.array(cnts)/2).astype(np.int16) )
#cnts = cnts/2
cnts = [(i/ 3).astype(np.int32) for i in cnts]
cnts_org = []
#print(cnts,'cnts')
for i in range(len(cnts)):
img_copy = np.zeros(img.shape)
img_copy = cv2.fillPoly(img_copy, pts=[cnts[i]], color=(1, 1, 1))
# plt.imshow(img_copy)
# plt.show()
# print(img.shape,'img')
img_copy = rotation_image_new(img_copy, -slope_first)
##print(img_copy.shape,'img_copy')
# plt.imshow(img_copy)
# plt.show()
img_copy = img_copy.astype(np.uint8)
imgray = cv2.cvtColor(img_copy, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(imgray, 0, 255, 0)
cont_int, _ = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
cont_int[0][:, 0, 0] = cont_int[0][:, 0, 0] + np.abs(img_copy.shape[1] - img.shape[1])
cont_int[0][:, 0, 1] = cont_int[0][:, 0, 1] + np.abs(img_copy.shape[0] - img.shape[0])
# print(np.shape(cont_int[0]))
cnts_org.append(cont_int[0]*3)
# sys.exit()
# self.y_shift = np.abs(img_copy.shape[0] - img.shape[0])
# self.x_shift = np.abs(img_copy.shape[1] - img.shape[1])
return cnts_org return cnts_org
def return_contours_of_interested_textline(region_pre_p, pixel): def return_contours_of_interested_textline(region_pre_p, pixel):