qurator-spk/eynollah
1
0
Fork 0
mirror of https://github.com/qurator-spk/eynollah.git synced 2025-10-15 19:09:58 +02:00
Code Issues Projects Releases Wiki Activity

utils_ocr: forgot to pass coordinate offsets

Browse source
This commit is contained in:
Robert Sachunsky 2025-10-08 14:56:14 +02:00
parent 839b7c4d84
commit 1d4815b48f
2 changed files with 20 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

24
src/eynollah/eynollah.py
View file

@ -4265,8 +4265,8 @@ class Eynollah:
if self.ocr and not self.tr: if self.ocr and not self.tr:
gc.collect() gc.collect()
ocr_all_textlines = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines = return_rnn_cnn_ocr_of_given_textlines(
image_page, all_found_textline_polygons, self.prediction_model, image_page, all_found_textline_polygons, np.zeros((len(all_found_textline_polygons), 4)),
self.b_s_ocr, self.num_to_char, textline_light=True) self.prediction_model, self.b_s_ocr, self.num_to_char, textline_light=True)
else: else:
ocr_all_textlines = None ocr_all_textlines = None
@ -4756,36 +4756,36 @@ class Eynollah:
if len(all_found_textline_polygons)>0: if len(all_found_textline_polygons)>0:
ocr_all_textlines = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines = return_rnn_cnn_ocr_of_given_textlines(
image_page, all_found_textline_polygons, self.prediction_model, image_page, all_found_textline_polygons, all_box_coord,
self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line) self.prediction_model, self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line)
else: else:
ocr_all_textlines = None ocr_all_textlines = None
if all_found_textline_polygons_marginals_left and len(all_found_textline_polygons_marginals_left)>0: if all_found_textline_polygons_marginals_left and len(all_found_textline_polygons_marginals_left)>0:
ocr_all_textlines_marginals_left = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines_marginals_left = return_rnn_cnn_ocr_of_given_textlines(
image_page, all_found_textline_polygons_marginals_left, self.prediction_model, image_page, all_found_textline_polygons_marginals_left, all_box_coord_marginals_left,
self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line) self.prediction_model, self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line)
else: else:
ocr_all_textlines_marginals_left = None ocr_all_textlines_marginals_left = None
if all_found_textline_polygons_marginals_right and len(all_found_textline_polygons_marginals_right)>0: if all_found_textline_polygons_marginals_right and len(all_found_textline_polygons_marginals_right)>0:
ocr_all_textlines_marginals_right = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines_marginals_right = return_rnn_cnn_ocr_of_given_textlines(
image_page, all_found_textline_polygons_marginals_right, self.prediction_model, image_page, all_found_textline_polygons_marginals_right, all_box_coord_marginals_right,
self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line) self.prediction_model, self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line)
else: else:
ocr_all_textlines_marginals_right = None ocr_all_textlines_marginals_right = None
if all_found_textline_polygons_h and len(all_found_textline_polygons)>0: if all_found_textline_polygons_h and len(all_found_textline_polygons)>0:
ocr_all_textlines_h = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines_h = return_rnn_cnn_ocr_of_given_textlines(
image_page, all_found_textline_polygons_h, self.prediction_model, image_page, all_found_textline_polygons_h, all_box_coord_h,
self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line) self.prediction_model, self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line)
else: else:
ocr_all_textlines_h = None ocr_all_textlines_h = None
if polygons_of_drop_capitals and len(polygons_of_drop_capitals)>0: if polygons_of_drop_capitals and len(polygons_of_drop_capitals)>0:
ocr_all_textlines_drop = return_rnn_cnn_ocr_of_given_textlines( ocr_all_textlines_drop = return_rnn_cnn_ocr_of_given_textlines(
image_page, polygons_of_drop_capitals, self.prediction_model, image_page, polygons_of_drop_capitals, np.zeros((len(polygons_of_drop_capitals), 4)),
self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line) self.prediction_model, self.b_s_ocr, self.num_to_char, self.textline_light, self.curved_line)
else: else:
ocr_all_textlines_drop = None ocr_all_textlines_drop = None

10
src/eynollah/utils/utils_ocr.py
View file

@ -1,13 +1,17 @@
import math
import copy
import numpy as np import numpy as np
import cv2 import cv2
import tensorflow as tf import tensorflow as tf
from scipy.signal import find_peaks from scipy.signal import find_peaks
from scipy.ndimage import gaussian_filter1d from scipy.ndimage import gaussian_filter1d
import math
from PIL import Image, ImageDraw, ImageFont from PIL import Image, ImageDraw, ImageFont
from Bio import pairwise2 from Bio import pairwise2
from .resize import resize_image from .resize import resize_image
def decode_batch_predictions(pred, num_to_char, max_len = 128): def decode_batch_predictions(pred, num_to_char, max_len = 128):
# input_len is the product of the batch size and the # input_len is the product of the batch size and the
# number of time steps. # number of time steps.
@ -370,7 +374,9 @@ def return_textline_contour_with_added_box_coordinate(textline_contour, box_ind
return textline_contour return textline_contour
def return_rnn_cnn_ocr_of_given_textlines(image, all_found_textline_polygons, def return_rnn_cnn_ocr_of_given_textlines(image,
all_found_textline_polygons,
all_box_coord,
prediction_model, prediction_model,
b_s_ocr, num_to_char, b_s_ocr, num_to_char,
textline_light=False, textline_light=False,