def create_style_dataset(output_dir, input_dir, b_dir):
if not os.path.exists(output_dir):
os.makedirs(output_dir)
src_paths = []
dst_paths = []
skipped = 0
for src_path in im.find(input_dir):
name, _ = os.path.splitext(os.path.basename(src_path))
dst_path = os.path.join(output_dir, name + ".png")
if os.path.exists(dst_path):
skipped += 1
else:
src_paths.append(src_path)
dst_paths.append(dst_path)
print("skipping %d files that already exist" % skipped)
global total
total = len(src_paths)
print("processing %d files" % total)
global start
start = time.time()
with tf.compat.v1.Session() as sess:
for src_path, dst_path in zip(src_paths, dst_paths):
src = im.load(src_path)
dst = combine(src, src_path, b_dir)
im.save(dst, dst_path)
complete()
def combine(pairing, A_path, B_path, dst_path):
# clear output path
if os.path.exists(dst_path):
shutil.rmtree(dst_path)
os.mkdir(dst_path)
with tf.Session() as sess:
for imgA_name, imgB_name in pairing.items():
imgA = im.load(os.path.join(A_path, imgA_name))
imgB = im.load(os.path.join(B_path, imgB_name))
# make sure that dimensions are correct
height, width, _ = imgA.shape
if height != imgB.shape[0] or width != imgB.shape[1]:
raise Exception("differing sizes")
# convert both images to RGB if necessary
if imgA.shape[2] == 1:
imgA = im.grayscale_to_rgb(images=imgA)
if imgB.shape[2] == 1:
imgB = im.grayscale_to_rgb(images=imgB)
# remove alpha channel
if imgA.shape[2] == 4:
imgA = imgA[:, :, :3]
if imgB.shape[2] == 4:
imgB = imgB[:, :, :3]
imgC = np.concatenate([imgA, imgB], axis=1)
im.save(imgC, os.path.join(dst_path, imgA_name))
def process(src_path, dst_path, image_dir):
global total_count
total_count += 1
src = im.load(src_path)
if args.operation == "combine":
dst = combine(src, src_path)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def process(src_path, dst_path):
src = im.load(src_path)
if a.operation == "grayscale":
dst = grayscale(src)
elif a.operation == "resize":
dst = resize(src)
elif a.operation == "blank":
dst = blank(src)
elif a.operation == "combine":
dst = combine(src, src_path)
elif a.operation == "edges":
dst = edges(src)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def process(src_path, dst_path):
src = im.load(src_path)
if a.operation == "grayscale":
dst = grayscale(src)
elif a.operation == "resize":
dst = resize(src)
elif a.operation == "blank":
dst = blank(src)
elif a.operation == "combine":
dst = combine(src, src_path)
elif a.operation == "edges":
dst = edges(src)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def main():
if not os.path.exists(a.output_dir):
os.makedirs(a.output_dir)
src_paths = []
dst_paths = []
for src_path in im.find(a.input_dir):
name, _ = os.path.splitext(os.path.basename(src_path))
dst_path = os.path.join(a.output_dir, name + ".png")
src_paths.append(src_path)
dst_paths.append(dst_path)
with tf.Session() as sess:
for src_path, dst_path in zip(src_paths, dst_paths):
src = im.load(src_path)
dst = combine(src, src_path)
im.save(dst, dst_path)
def process(src_path, dst_path):
src = im.load(src_path)
if a.operation == "grayscale": #灰度化
dst = grayscale(src)
elif a.operation == "resize": #缩放至统一尺寸
dst = resize(src)
elif a.operation == "blank": #中间留白
dst = blank(src)
elif a.operation == "combine": #结合图片
dst = combine(src, src_path)
elif a.operation == "edges": #检测边缘
dst = edges(src)
elif a.operation == "blur": #模糊
dst = blur(src)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def process(src_path, dst_path): #通过Process()函数，根据命令行参数，对image进行相应处理
src = im.load(src_path)
if a.operation == "grayscale":
dst = grayscale(src)
elif a.operation == "resize":
dst = resize(src)
elif a.operation == "blank":
dst = blank(src)
elif a.operation == "combine":
dst = combine(src, src_path)
elif a.operation == "edges":
dst = edges(src)
elif a.operation == "blur":
dst = blur(src)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def process(src_path, dst_path):
src = im.load(src_path)
im.save(src, "/data12/liyh/tools/test_output/src.png")
if a.operation == "grayscale":
dst = grayscale(src)
elif a.operation == "resize":
dst = resize(src)
elif a.operation == "blank":
dst = blank(src)
elif a.operation == "combine":
dst = combine(src, src_path)
elif a.operation == "edges":
dst = edges(src)
elif a.operation == "combine_color_binary":
dst = combine_color_binary(src, src_path)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
def process(src_path, dst_path, label_dict, labels, labels_csv, image_dir):
global index
global total_count
# 입력받은 하나의 한글 이미지 경로를 가지고 해당 한글 이미지를 읽음
total_count += 1
src = im.load(src_path)
# 명령어에서 인자로 입력받은 연산이 combine이라면 combine 수행
if args.operation == "combine":
dst = combine(src, src_path)
else:
raise Exception("invalid operation")
# combine을 수행한 결과를 파일로 저장함
im.save(dst, dst_path)
# 저장한 combine 결과 이미지를 csv 파일에 레이블을 맵핑하여 저장
file_string = '{}.png'.format(total_count)
file_path = os.path.join(image_dir, file_string)
character = list(label_dict.keys())[labels[index]]
labels_csv.write(u'{},{}\n'.format(file_path, character))
index += 1
def process(src_path, dst_path):
src = im.load(src_path)
if a.operation == "edges":
if a.crop:
name = dst_path.split("/")[-1]
src, dst = edges(src)
im.save(src, os.path.join(a.crop_dir, name))
else:
dst = edges(src)
elif a.operation == "grayscale":
dst = grayscale(src)
elif a.operation == "resize":
dst = resize(src)
elif a.operation == "blank":
dst = blank(src)
elif a.operation == "combine":
dst = combine(src, src_path)
elif a.operation == "skeletonize":
dst = skeletonize_edge(src)
else:
raise Exception("invalid operation")
im.save(dst, dst_path)
# convert both images to RGB if necessary
if src1.shape[2] == 1:
src1 = im.grayscale_to_rgb(images=src1)
src2 = im.grayscale_to_rgb(images=src2)
# remove alpha channel
if src2.shape[2] == 4:
src1 = src1[:, :, :3]
src2 = src2[:, :, :3]
im1 = src1[:, :width_cutoff]
im2 = src2[:, width_cutoff:]
im_dest = np.concatenate([im1, im2], axis=1)
return im_dest
with tf.Session() as sess:
for src_path1, src_path2, dst_path_new in zip(src_paths1, src_paths2,
dst_paths):
srcA = im.load(src_path1)
srcB = im.load(src_path2)
new_im = split(srcA, srcB)
#print(src_path1)
#print(src_path2)
#print(dst_path_new)
test = 1
im.save(new_im, dst_path_new)
def main():
for output_type in output_types:
full_subfolder_path = os.path.join(a.output_dir, output_type)
if not os.path.exists(full_subfolder_path):
os.makedirs(full_subfolder_path)
seg_input_files = im.findContainingSubtext(a.input_dir, 'sample_inputs')
gen_input_files = im.findContainingSubtext(a.input_dir, 'sample_outputs')
output_paths = {}
seg_src_paths = {}
gen_src_paths = {}
skipped = 0
for seg_input_file, gen_input_file in zip(seg_input_files,
gen_input_files):
name, _ = os.path.splitext(os.path.basename(seg_input_file))
name = name.split('_')[0]
output_paths[name] = {}
name_out_dict = {}
output_file_exists_for_name = False
for output_type in output_types:
output_path = os.path.join(a.output_dir, output_type)
output_path = os.path.join(output_path, name + '.png')
if os.path.exists(output_path):
output_file_exists_for_name = True
break
name_out_dict[output_type] = output_path
if not output_file_exists_for_name:
output_paths[name] = name_out_dict
seg_src_paths[name] = seg_input_file
gen_src_paths[name] = gen_input_file
else:
skipped += 1
print("skipping %d files that already exist" % skipped)
global total
total = len(output_paths)
print("processing %d files" % total)
global start
start = time.time()
skin_color = getSkinColor()
background_color = getBackgroundColor()
positive_colors = getPositiveColors()
upscale_interpolation = getUpscaleInterpolation()
with tf.Session() as sess:
for name, name_out_dict in output_paths.iteritems():
print name
print name_out_dict
gen_image = im.load(gen_src_paths[name])
seg_image = im.load(seg_src_paths[name])
#Save inputs
im.save(seg_image, name_out_dict['humanseginputs'])
im.save(gen_image, name_out_dict['humangeninputs'])
#Crop reference
crop_reference = getFaceCropReference(seg_image, skin_color)
#Cropped face segmentation
temp_image_seg = crop(seg_image, crop_reference)
temp_image_seg = addBackground(temp_image_seg, skin_color,
positive_colors, background_color)
im.save(temp_image_seg, name_out_dict['croppedsegfaces'])
#Resize face segmentation
temp_image_seg = resize(
temp_image_seg,
tfUpscaleInterpolationType=upscale_interpolation)
im.save(temp_image_seg, name_out_dict['resizedsegfaces'])
#Cropped face generated
temp_image_gen = crop(gen_image, crop_reference)
im.save(temp_image_gen, name_out_dict['croppedgenfaces'])
#Resize face generated
temp_image_gen = resize(
temp_image_gen,
tfUpscaleInterpolationType=upscale_interpolation)
im.save(temp_image_gen, name_out_dict['resizedgenfaces'])
complete()
#Combine resized seg and gen images to input face module
inputDirA = os.path.join(a.output_dir, 'resizedsegfaces')
inputDirB = os.path.join(a.output_dir, 'resizedgenfaces')
outputDir = os.path.join(a.output_dir, 'combinedseggenfaces')
combineImagePairs(inputDirA, inputDirB, outputDir)
#Now lets generate face images from extracted face segmentation maps
inputDir = os.path.join(a.output_dir, 'combinedseggenfaces')
outputDir = os.path.join(a.output_dir, 'generatedfaces')
runFacePix2PixModule(inputDir, outputDir)
with tf.Session() as sess:
for name, name_out_dict in output_paths.iteritems():
gen_image = im.load(name_out_dict['humangeninputs'])
seg_image = im.load(name_out_dict['humanseginputs'])
#Crop reference
crop_reference = getFaceCropReference(seg_image, skin_color)
#Get crop reference point and size
[rmin, rmax, cmin, cmax] = cropRectangle(crop_reference)
width = rmax - rmin
height = cmax - cmin
#Load generated face
gen_face_path = os.path.join(a.output_dir, 'generatedfaces')
gen_face_path = os.path.join(gen_face_path, 'images')
gen_face_path = os.path.join(gen_face_path, name + '-outputs.png')
gen_face_image = im.load(gen_face_path)
gen_face_image = downscaleToSize(gen_face_image, [width, height])
updated_full_body_image = updateFaceResults(
gen_image, gen_face_image, rmin, rmax, cmin, cmax,
crop_reference)
im.save(updated_full_body_image, name_out_dict['mergedgenoutputs'])
#Combine original generated full body images and face updated generated images
inputDirA = os.path.join(a.output_dir, 'humangeninputs')
inputDirB = os.path.join(a.output_dir, 'mergedgenoutputs')
outputDir = os.path.join(a.output_dir, 'finalcomparisonimages')
combineImagePairs(inputDirA, inputDirB, outputDir)
def main():
if not os.path.exists(a.output_dir_labels):
os.makedirs(a.output_dir_labels)
if not os.path.exists(output_train_directory_labels):
os.makedirs(output_train_directory_labels)
if not os.path.exists(output_test_directory_labels):
os.makedirs(output_test_directory_labels)
if not os.path.exists(output_val_directory_labels):
os.makedirs(output_val_directory_labels)
processInputImages = a.resize or a.crop
if not os.path.exists(a.output_dir_images) and processInputImages:
os.makedirs(a.output_dir_images)
if not os.path.exists(output_train_directory_images) and processInputImages:
os.makedirs(output_train_directory_images)
if not os.path.exists(output_test_directory_images) and processInputImages:
os.makedirs(output_test_directory_images)
if not os.path.exists(output_val_directory_images) and processInputImages:
os.makedirs(output_val_directory_images)
#cropped images directory
splits = ['train', 'test', 'val']
src_paths = []
dst_paths_labels = []
dst_paths_images = []
skipped = 0
for split in splits:
split_folder = os.path.join(a.input_dir, split)
for src_path in im.find(split_folder):
name, _ = os.path.splitext(os.path.basename(src_path))
dst_path_label = os.path.join(a.output_dir_labels, split)
dst_path_label = os.path.join(dst_path_label, name + ".png")
dst_path_image = os.path.join(a.output_dir_images, split)
dst_path_image = os.path.join(dst_path_image, name + ".png")
if os.path.exists(dst_path_label) or os.path.exists(dst_path_image):
skipped += 1
else:
src_paths.append(src_path)
dst_paths_labels.append(dst_path_label)
dst_paths_images.append(dst_path_image)
print("skipping %d files that already exist" % skipped)
global total
total = len(src_paths)
print("processing %d files" % total)
global start
start = time.time()
if a.workers == 1:
with tf.Session() as sess:
for src_path, dst_path_label, dst_path_image in zip(src_paths, dst_paths_labels, dst_paths_images):
name, _ = os.path.splitext(os.path.basename(src_path))
print 'Name: ' + name
label_folder = os.path.join(a.label_images_dir, name)
label_image_paths = getLabelImages(label_folder)
print label_image_paths
color_dict = getLabelToColorDictionary()
label_img = getLabelImage(label_image_paths, color_dict)
if processInputImages:
processedImage = im.load(src_path)
if a.crop:
crop_reference = getCropReference(label_image_paths)
processedImage = crop(processedImage, crop_reference)
label_img = crop(label_img, crop_reference)
if a.resize:
processedImage = resize(processedImage)
label_img = resize(label_img)
im.save(processedImage, dst_path_image)
im.save(label_img, dst_path_label)
complete()