def detect_img_folder(img_folder, out_folder, yolo):
mkdir_if_not_exist(out_folder)
path_list, name_list = traverse_dir_files(img_folder)
print_info('图片数: %s' % len(path_list))
_, imgs_names = traverse_dir_files(out_folder)
count = 0
for path, name in zip(path_list, name_list):
if path.endswith('.gif'):
continue
out_name = name + '.d.jpg'
if out_name in imgs_names:
print_info('已检测: %s' % name)
continue
print_info('检测图片: %s' % name)
try:
image = Image.open(path)
out_file = os.path.join(ROOT_DIR, 'face', 'yolov3', 'output_data',
'logAll_res.txt')
r_image = yolo.detect_image(image, ('logAll/' + name), out_file)
r_image.save(os.path.join(out_folder, name + '.d.jpg'))
except Exception as e:
print(e)
pass
count += 1
if count % 100 == 0:
print_info('已检测: %s' % count)
yolo.close_session()
def format_img_and_anno(img_folder):
"""
格式化输出。图片和标注文件夹
:param img_folder: 图片文件夹
:return:
"""
file_paths, file_names = traverse_dir_files(img_folder)
img_dict = dict() # 将标注和图片路径，生成一个字典
for file_path, file_name in zip(file_paths, file_names):
if file_name.endswith('.jpg'):
name = file_name.replace('.jpg', '')
if name not in img_dict:
img_dict[name] = (None, None)
(img_p, anno_p) = img_dict[name]
img_dict[name] = (file_path, anno_p)
if file_name.endswith('.xml'):
name = file_name.replace('.xml', '')
if name not in img_dict:
img_dict[name] = (None, None)
(img_p, anno_p) = img_dict[name]
img_dict[name] = (img_p, file_path)
print_info('图片数: {}'.format(len(img_dict.keys())))
return img_dict
def data_processor_testV3():
dataset_dir = os.path.join(DATASET_DIR, 's2a4zsV4')
# person_path = "/Users/wangchenlong/Downloads/seeprettyface_asian_stars"
person_path = "/Users/wangchenlong/Downloads/SCUT-FBP5500_v2/Images"
paths_list, names_list = traverse_dir_files(person_path)
trainA_dir = os.path.join(dataset_dir, 'trainA')
testA_dir = os.path.join(dataset_dir, 'testA')
mkdir_if_not_exist(trainA_dir)
mkdir_if_not_exist(testA_dir)
train_size = 5000
test_size = 100
print_size = 100
count = 0
random.shuffle(paths_list)
for path in paths_list:
img = cv2.imread(path)
img = cv2.resize(img, (256, 256))
if count < train_size:
file_name = os.path.join(trainA_dir,
u"c_{:04d}.jpg".format(count + 1))
cv2.imwrite(file_name, img)
else:
file_name = os.path.join(testA_dir,
u"c_{:04d}.jpg".format(count + 1))
cv2.imwrite(file_name, img)
count += 1
if count % print_size == 0:
print(u'[Info] run count: {}'.format(count))
if count == train_size + test_size:
break
print('[Info] 数据处理完成')
def write_frames_to_vid(self, frames_folder, out_vid_path):
paths_list, names_list = traverse_dir_files(frames_folder)
img_list = []
for path, name in zip(paths_list, names_list):
img = cv2.imread(path)
img_list.append(img)
# fps: 29, h: 1280, w: 720
write_video(out_vid_path, img_list, 29, 1280, 720)
def get_livevqc_index(feature_dir):
from utils.project_utils import traverse_dir_files
paths_list, names_list = traverse_dir_files(feature_dir, ext='.npy')
res_names = set()
for name in names_list:
res_names.add(name.split('_')[0])
res_names = sorted(list(res_names))
return res_names
def save_video():
img_dir = os.path.join(DATA_DIR, 'frames')
paths_list, names_list = traverse_dir_files(img_dir)
frame_list = []
for name, path in zip(names_list, paths_list):
frame = cv.imread(path)
frame_list.append(frame)
from_video = "normal_video.mp4"
cap, n_frame, fps, h, w = init_vid(from_video)
video_path = "norm.out.mp4"
write_video(video_path, frame_list, fps, h, w)
def folder_test():
img_dir = os.path.join(IMGS_DIR, 'tests')
out_dir = os.path.join(IMGS_DIR, 'tests-out')
paths_list, names_list = traverse_dir_files(img_dir)
gp = GazePredicter()
for img_path, name in zip(paths_list, names_list):
print('[Info] 处理图像: {}'.format(name))
face_dict = gp.predict_path(img_path)
img_op = face_dict['img_draw']
out_path = os.path.join(out_dir, name + ".out.jpg")
cv2.imwrite(out_path, img_op)
def process_dataset(self):
"""
处理数据集
"""
c_paths_list, c_names_list = traverse_dir_files(self.cartoons_path)
p_paths_list, p_names_list = traverse_dir_files(self.persons_path)
random.seed(47)
random.shuffle(c_paths_list)
random.shuffle(p_paths_list)
train_size = 1500 # 训练集量
test_size = 100 # 测试集量
print_size = 100
count = 0
train_person_dir = os.path.join(ROOT_DIR, 'dataset', 's2a4zsV1',
'trainA')
test_person_dir = os.path.join(ROOT_DIR, 'dataset', 's2a4zsV1',
'testA')
mkdir_if_not_exist(train_person_dir)
mkdir_if_not_exist(test_person_dir)
print('[Info] 真人样本总数: {}'.format(len(p_paths_list)))
for p_path in p_paths_list:
try:
p_img = cv2.imread(p_path)
p_img = cv2.resize(p_img, (256, 256))
if count < train_size:
p_file_name = os.path.join(
train_person_dir, u"p_{:04d}.jpg".format(count + 1))
else:
p_file_name = os.path.join(
test_person_dir, u"p_{:04d}.jpg".format(count + 1))
cv2.imwrite(p_file_name, p_img)
count += 1
except Exception as e:
print('[Error] error {}'.format(e))
continue
if count % print_size == 0:
print(u'[Info] run count: {}'.format(count))
if count == train_size + test_size:
break
train_cartoon_dir = os.path.join(ROOT_DIR, 'dataset', 's2a4zsV1',
'trainB')
test_cartoon_dir = os.path.join(ROOT_DIR, 'dataset', 's2a4zsV1',
'testB')
mkdir_if_not_exist(train_cartoon_dir)
mkdir_if_not_exist(test_cartoon_dir)
count = 0
print('[Info] 卡通样本总数: {}'.format(len(c_paths_list)))
for c_path in c_paths_list:
try:
c_img = cv2.imread(c_path)
c_img = cv2.resize(c_img, (256, 256))
if count < train_size:
c_file_name = os.path.join(
train_cartoon_dir, u"c_{:04d}.jpg".format(count + 1))
cv2.imwrite(c_file_name, c_img)
else:
c_file_name = os.path.join(
test_cartoon_dir, u"c_{:04d}.jpg".format(count + 1))
cv2.imwrite(c_file_name, c_img)
count += 1
except Exception as e:
print('[Error] error {}'.format(e))
continue
if count % print_size == 0:
print(u'[Info] run count: {}'.format(count))
if count == train_size + test_size:
break
print('[Info] 数据处理完成')