Merge 4abd5b281e into 39ef3fd7bb
Alexander Pacha
2 years ago
committed by
GitHub
commit
bf954a9319
@ -1,262 +1,187 @@
|
|||||||
"""
|
import argparse
|
||||||
Tool to load model and binarize a given image.
|
import gc
|
||||||
"""
|
import itertools
|
||||||
|
import math
|
||||||
|
import os
|
||||||
import sys
|
import sys
|
||||||
from glob import glob
|
from pathlib import Path
|
||||||
from os import environ, devnull
|
from typing import Union, List, Tuple, Any
|
||||||
from os.path import join
|
|
||||||
from warnings import catch_warnings, simplefilter
|
|
||||||
|
|
||||||
import numpy as np
|
|
||||||
from PIL import Image
|
|
||||||
import cv2
|
import cv2
|
||||||
environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
|
import numpy as np
|
||||||
|
|
||||||
|
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
|
||||||
stderr = sys.stderr
|
stderr = sys.stderr
|
||||||
sys.stderr = open(devnull, 'w')
|
sys.stderr = open(os.devnull, 'w')
|
||||||
import tensorflow as tf
|
import tensorflow as tf
|
||||||
from tensorflow.keras.models import load_model
|
from tensorflow.python.keras.saving.save import load_model
|
||||||
from tensorflow.python.keras import backend as tensorflow_backend
|
|
||||||
sys.stderr = stderr
|
sys.stderr = stderr
|
||||||
|
|
||||||
|
from mpire import WorkerPool
|
||||||
|
from mpire.utils import make_single_arguments
|
||||||
|
|
||||||
import logging
|
|
||||||
|
|
||||||
def resize_image(img_in, input_height, input_width):
|
|
||||||
return cv2.resize(img_in, (input_width, input_height), interpolation=cv2.INTER_NEAREST)
|
|
||||||
|
|
||||||
class SbbBinarizer:
|
class SbbBinarizer:
|
||||||
|
def __init__(self) -> None:
|
||||||
|
super().__init__()
|
||||||
|
self.models: List[Tuple[Any, int, int, int]] = []
|
||||||
|
|
||||||
|
def load_model(self, model_dir: Union[str, Path]):
|
||||||
|
model_dir = Path(model_dir)
|
||||||
|
model_paths = list(model_dir.glob('*.h5')) or list(model_dir.glob('*/'))
|
||||||
|
for path in model_paths:
|
||||||
|
model = load_model(str(path.absolute()), compile=False)
|
||||||
|
height = model.layers[len(model.layers) - 1].output_shape[1]
|
||||||
|
width = model.layers[len(model.layers) - 1].output_shape[2]
|
||||||
|
classes = model.layers[len(model.layers) - 1].output_shape[3]
|
||||||
|
self.models.append((model, height, width, classes))
|
||||||
|
|
||||||
|
def binarize_image_file(self, image_path: Path, save_path: Path):
|
||||||
|
if not image_path.exists():
|
||||||
|
raise ValueError(f"Image not found: {str(image_path)}")
|
||||||
|
|
||||||
|
# noinspection PyUnresolvedReferences
|
||||||
|
img = cv2.imread(str(image_path))
|
||||||
|
|
||||||
|
full_image = self.binarize_image(img)
|
||||||
|
|
||||||
|
Path(save_path).parent.mkdir(parents=True, exist_ok=True)
|
||||||
|
# noinspection PyUnresolvedReferences
|
||||||
|
cv2.imwrite(str(save_path), full_image)
|
||||||
|
|
||||||
|
def binarize_image(self, img: np.ndarray) -> np.ndarray:
|
||||||
|
img_last = False
|
||||||
|
for model, model_height, model_width, _ in self.models:
|
||||||
|
img_res = self.binarize_image_by_model(img, model, model_height, model_width)
|
||||||
|
img_last = img_last + (img_res == 0)
|
||||||
|
img_last = (~img_last).astype(np.uint8) * 255
|
||||||
|
return img_last
|
||||||
|
|
||||||
def __init__(self, model_dir, logger=None):
|
def binarize_image_by_model(self, img: np.ndarray, model: Any, model_height: int, model_width: int) -> np.ndarray:
|
||||||
self.model_dir = model_dir
|
# Padded images must be multiples of model size
|
||||||
self.log = logger if logger else logging.getLogger('SbbBinarizer')
|
original_image_height, original_image_width, image_channels = img.shape
|
||||||
|
|
||||||
self.start_new_session()
|
padded_image_height = math.ceil(original_image_height / model_height) * model_height
|
||||||
|
padded_image_width = math.ceil(original_image_width / model_width) * model_width
|
||||||
self.model_files = glob('%s/*.h5' % self.model_dir)
|
padded_image = np.zeros((padded_image_height, padded_image_width, image_channels))
|
||||||
if not self.model_files:
|
padded_image[0:original_image_height, 0:original_image_width, :] = img[:, :, :]
|
||||||
self.model_files = glob('%s/*/' % self.model_dir)
|
|
||||||
if not self.model_files:
|
image_batch = np.expand_dims(padded_image, 0) # Create the batch dimension
|
||||||
raise ValueError(f"No models found in {self.model_dir}")
|
patches = tf.image.extract_patches(
|
||||||
|
images=image_batch,
|
||||||
self.models = []
|
sizes=[1, model_height, model_width, 1],
|
||||||
for model_file in self.model_files:
|
strides=[1, model_height, model_width, 1],
|
||||||
self.models.append(self.load_model(model_file))
|
rates=[1, 1, 1, 1],
|
||||||
|
padding='SAME'
|
||||||
def start_new_session(self):
|
)
|
||||||
config = tf.compat.v1.ConfigProto()
|
|
||||||
config.gpu_options.allow_growth = True
|
number_of_horizontal_patches = patches.shape[1]
|
||||||
|
number_of_vertical_patches = patches.shape[2]
|
||||||
self.session = tf.compat.v1.Session(config=config) # tf.InteractiveSession()
|
total_number_of_patches = number_of_horizontal_patches * number_of_vertical_patches
|
||||||
tensorflow_backend.set_session(self.session)
|
target_shape = (total_number_of_patches, model_height, model_width, image_channels)
|
||||||
|
# Squeeze all image patches (n, m, width, height, channels) into a single big batch (b, width, height, channels)
|
||||||
def end_session(self):
|
image_patches = tf.reshape(patches, target_shape)
|
||||||
tensorflow_backend.clear_session()
|
# Normalize the image to values between 0.0 - 1.0
|
||||||
self.session.close()
|
image_patches = image_patches / float(255.0)
|
||||||
del self.session
|
|
||||||
|
predicted_patches = model.predict(image_patches, verbose=0)
|
||||||
def load_model(self, model_name):
|
# We have to manually call garbage collection and clear_session here to avoid memory leaks.
|
||||||
model = load_model(model_name, compile=False)
|
# Taken from https://medium.com/dive-into-ml-ai/dealing-with-memory-leak-issue-in-keras-model-training-e703907a6501
|
||||||
model_height = model.layers[len(model.layers)-1].output_shape[1]
|
gc.collect()
|
||||||
model_width = model.layers[len(model.layers)-1].output_shape[2]
|
tf.keras.backend.clear_session()
|
||||||
n_classes = model.layers[len(model.layers)-1].output_shape[3]
|
|
||||||
return model, model_height, model_width, n_classes
|
# The result is a white-on-black image that needs to be inverted to be displayed as black-on-white image
|
||||||
|
# We do this by converting the binary values to a boolean numpy-array and then inverting the values
|
||||||
def predict(self, model_in, img):
|
black_on_white_patches = np.invert(np.argmax(predicted_patches, axis=3).astype(bool))
|
||||||
tensorflow_backend.set_session(self.session)
|
# cv2 can't export a boolean numpy array into a black-and-white PNG image, so we have to convert it to uint8 (grayscale) values
|
||||||
model, model_height, model_width, n_classes = model_in
|
grayscale_patches = black_on_white_patches.astype(np.uint8) * 255
|
||||||
|
full_image_with_padding = self._patches_to_image(
|
||||||
img_org_h = img.shape[0]
|
grayscale_patches,
|
||||||
img_org_w = img.shape[1]
|
padded_image_height,
|
||||||
|
padded_image_width,
|
||||||
if img.shape[0] < model_height and img.shape[1] >= model_width:
|
model_height,
|
||||||
img_padded = np.zeros(( model_height, img.shape[1], img.shape[2] ))
|
model_width
|
||||||
|
)
|
||||||
index_start_h = int( abs( img.shape[0] - model_height) /2.)
|
full_image = full_image_with_padding[0:original_image_height, 0:original_image_width]
|
||||||
index_start_w = 0
|
return full_image
|
||||||
|
|
||||||
img_padded [ index_start_h: index_start_h+img.shape[0], :, : ] = img[:,:,:]
|
def _patches_to_image(self, patches: np.ndarray, image_height: int, image_width: int, patch_height: int, patch_width: int):
|
||||||
|
height = math.ceil(image_height / patch_height) * patch_height
|
||||||
elif img.shape[0] >= model_height and img.shape[1] < model_width:
|
width = math.ceil(image_width / patch_width) * patch_width
|
||||||
img_padded = np.zeros(( img.shape[0], model_width, img.shape[2] ))
|
|
||||||
|
image_reshaped = np.reshape(
|
||||||
index_start_h = 0
|
np.squeeze(patches),
|
||||||
index_start_w = int( abs( img.shape[1] - model_width) /2.)
|
[height // patch_height, width // patch_width, patch_height, patch_width]
|
||||||
|
)
|
||||||
img_padded [ :, index_start_w: index_start_w+img.shape[1], : ] = img[:,:,:]
|
image_transposed = np.transpose(a=image_reshaped, axes=[0, 2, 1, 3])
|
||||||
|
image_resized = np.reshape(image_transposed, [height, width])
|
||||||
|
return image_resized
|
||||||
elif img.shape[0] < model_height and img.shape[1] < model_width:
|
|
||||||
img_padded = np.zeros(( model_height, model_width, img.shape[2] ))
|
|
||||||
|
def split_list_into_worker_batches(files: List[Any], number_of_workers: int) -> List[List[Any]]:
|
||||||
index_start_h = int( abs( img.shape[0] - model_height) /2.)
|
""" Splits any given list into batches for the specified number of workers and returns a list of lists. """
|
||||||
index_start_w = int( abs( img.shape[1] - model_width) /2.)
|
batches = []
|
||||||
|
batch_size = math.ceil(len(files) / number_of_workers)
|
||||||
img_padded [ index_start_h: index_start_h+img.shape[0], index_start_w: index_start_w+img.shape[1], : ] = img[:,:,:]
|
batch_start = 0
|
||||||
|
for i in range(1, number_of_workers + 1):
|
||||||
else:
|
batch_end = i * batch_size
|
||||||
index_start_h = 0
|
file_batch_to_delete = files[batch_start: batch_end]
|
||||||
index_start_w = 0
|
batches.append(file_batch_to_delete)
|
||||||
img_padded = np.copy(img)
|
batch_start = batch_end
|
||||||
|
return batches
|
||||||
|
|
||||||
img = np.copy(img_padded)
|
|
||||||
|
def batch_predict(input_data):
|
||||||
margin = int(0.1 * model_width)
|
model_dir, input_images, output_images, worker_number = input_data
|
||||||
|
print(f"Setting visible cuda devices to {str(worker_number)}")
|
||||||
width_mid = model_width - 2 * margin
|
# Each worker thread will be assigned only one of the available GPUs to allow multiprocessing across GPUs
|
||||||
height_mid = model_height - 2 * margin
|
os.environ["CUDA_VISIBLE_DEVICES"] = str(worker_number)
|
||||||
|
|
||||||
|
binarizer = SbbBinarizer()
|
||||||
img = img / float(255.0)
|
binarizer.load_model(model_dir)
|
||||||
|
|
||||||
img_h = img.shape[0]
|
for image_path, output_path in zip(input_images, output_images):
|
||||||
img_w = img.shape[1]
|
binarizer.binarize_image(image_path=image_path, save_path=output_path)
|
||||||
|
print(f"Binarized {image_path}")
|
||||||
prediction_true = np.zeros((img_h, img_w, 3))
|
|
||||||
mask_true = np.zeros((img_h, img_w))
|
|
||||||
nxf = img_w / float(width_mid)
|
if __name__ == '__main__':
|
||||||
nyf = img_h / float(height_mid)
|
parser = argparse.ArgumentParser()
|
||||||
|
parser.add_argument('-m', '--model_dir', default="model_2021_03_09", help="Path to the directory where the TF model resides or path to an h5 file.")
|
||||||
if nxf > int(nxf):
|
parser.add_argument('-i', '--input-path', required=True)
|
||||||
nxf = int(nxf) + 1
|
parser.add_argument('-o', '--output-path', required=True)
|
||||||
else:
|
args = parser.parse_args()
|
||||||
nxf = int(nxf)
|
|
||||||
|
input_path = Path(args.input_path)
|
||||||
if nyf > int(nyf):
|
output_path = Path(args.output_path)
|
||||||
nyf = int(nyf) + 1
|
model_directory = args.model_dir
|
||||||
else:
|
|
||||||
nyf = int(nyf)
|
if input_path.is_dir():
|
||||||
|
print(f"Enumerating all PNG files in {str(input_path)}")
|
||||||
for i in range(nxf):
|
all_input_images = list(input_path.rglob("*.png"))
|
||||||
for j in range(nyf):
|
print(f"Filtering images that have already been binarized in {str(output_path)}")
|
||||||
|
input_images = [i for i in all_input_images if not (output_path / (i.relative_to(input_path))).exists()]
|
||||||
if i == 0:
|
output_images = [output_path / (i.relative_to(input_path)) for i in input_images]
|
||||||
index_x_d = i * width_mid
|
input_images = [i for i in input_images]
|
||||||
index_x_u = index_x_d + model_width
|
|
||||||
elif i > 0:
|
print(f"Starting batch-binarization of {len(input_images)} images")
|
||||||
index_x_d = i * width_mid
|
|
||||||
index_x_u = index_x_d + model_width
|
number_of_gpus = len(tf.config.list_physical_devices('GPU'))
|
||||||
|
number_of_workers = max(1, number_of_gpus)
|
||||||
if j == 0:
|
image_batches = split_list_into_worker_batches(input_images, number_of_workers)
|
||||||
index_y_d = j * height_mid
|
output_batches = split_list_into_worker_batches(output_images, number_of_workers)
|
||||||
index_y_u = index_y_d + model_height
|
|
||||||
elif j > 0:
|
# Must use spawn to create completely new process that has its own resources to properly multiprocess across GPUs
|
||||||
index_y_d = j * height_mid
|
with WorkerPool(n_jobs=number_of_workers, start_method='spawn') as pool:
|
||||||
index_y_u = index_y_d + model_height
|
model_dirs = itertools.repeat(model_directory, len(image_batches))
|
||||||
|
input_data = zip(model_dirs, image_batches, output_batches, range(number_of_workers))
|
||||||
if index_x_u > img_w:
|
contents = pool.map_unordered(
|
||||||
index_x_u = img_w
|
batch_predict,
|
||||||
index_x_d = img_w - model_width
|
make_single_arguments(input_data),
|
||||||
if index_y_u > img_h:
|
iterable_len=number_of_workers,
|
||||||
index_y_u = img_h
|
progress_bar=False
|
||||||
index_y_d = img_h - model_height
|
)
|
||||||
|
|
||||||
img_patch = img[index_y_d:index_y_u, index_x_d:index_x_u, :]
|
|
||||||
|
|
||||||
label_p_pred = model.predict(img_patch.reshape(1, img_patch.shape[0], img_patch.shape[1], img_patch.shape[2]))
|
|
||||||
|
|
||||||
seg = np.argmax(label_p_pred, axis=3)[0]
|
|
||||||
|
|
||||||
seg_color = np.repeat(seg[:, :, np.newaxis], 3, axis=2)
|
|
||||||
|
|
||||||
if i == 0 and j == 0:
|
|
||||||
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]
|
|
||||||
|
|
||||||
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
|
|
||||||
|
|
||||||
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 = 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
|
|
||||||
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:
|
|
||||||
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]
|
|
||||||
|
|
||||||
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
|
|
||||||
|
|
||||||
elif i == nxf-1 and j == 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]
|
|
||||||
|
|
||||||
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
|
|
||||||
|
|
||||||
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 = 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
|
|
||||||
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:
|
|
||||||
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]
|
|
||||||
|
|
||||||
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
|
|
||||||
|
|
||||||
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 = 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
|
|
||||||
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:
|
|
||||||
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]
|
|
||||||
|
|
||||||
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
|
|
||||||
|
|
||||||
else:
|
else:
|
||||||
seg_color = seg_color[margin:seg_color.shape[0] - margin, margin:seg_color.shape[1] - margin, :]
|
binarizer = SbbBinarizer()
|
||||||
seg = seg[margin:seg.shape[0] - margin, margin:seg.shape[1] - margin]
|
binarizer.load_model(model_directory)
|
||||||
|
binarizer.binarize_image(image_path=input_path, save_path=output_path)
|
||||||
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 = prediction_true[index_start_h: index_start_h+img_org_h, index_start_w: index_start_w+img_org_w,:]
|
|
||||||
prediction_true = prediction_true.astype(np.uint8)
|
|
||||||
|
|
||||||
return prediction_true[:,:,0]
|
|
||||||
|
|
||||||
def run(self, image=None, image_path=None, save=None):
|
|
||||||
if (image is not None and image_path is not None) or \
|
|
||||||
(image is None and image_path is None):
|
|
||||||
raise ValueError("Must pass either a opencv2 image or an image_path")
|
|
||||||
if image_path is not None:
|
|
||||||
image = cv2.imread(image_path)
|
|
||||||
img_last = 0
|
|
||||||
for n, (model, model_file) in enumerate(zip(self.models, self.model_files)):
|
|
||||||
self.log.info('Predicting with model %s [%s/%s]' % (model_file, n + 1, len(self.model_files)))
|
|
||||||
|
|
||||||
res = self.predict(model, image)
|
|
||||||
|
|
||||||
img_fin = np.zeros((res.shape[0], res.shape[1], 3))
|
|
||||||
res[:, :][res[:, :] == 0] = 2
|
|
||||||
res = res - 1
|
|
||||||
res = res * 255
|
|
||||||
img_fin[:, :, 0] = res
|
|
||||||
img_fin[:, :, 1] = res
|
|
||||||
img_fin[:, :, 2] = res
|
|
||||||
|
|
||||||
img_fin = img_fin.astype(np.uint8)
|
|
||||||
img_fin = (res[:, :] == 0) * 255
|
|
||||||
img_last = img_last + img_fin
|
|
||||||
|
|
||||||
kernel = np.ones((5, 5), np.uint8)
|
|
||||||
img_last[:, :][img_last[:, :] > 0] = 255
|
|
||||||
img_last = (img_last[:, :] == 0) * 255
|
|
||||||
if save:
|
|
||||||
cv2.imwrite(save, img_last)
|
|
||||||
return img_last
|
|
||||||
|
|||||||
Loading…
Reference in New Issue