def run_detect(filename):
base_file_name = os.path.basename(filename)
saved_file_name = os.path.join(MASKED_DIR, base_file_name)
logging.info('Loading image: %s', base_file_name)
# Convert png with alpha channel with shape[2] == 4 into shape[2] ==3 RGB images
image = skimage.io.imread(filename)
if len(image.shape) > 2 and image.shape[2] == 4:
image = cv2.cvtColor(image, cv2.COLOR_BGRA2BGR)
logging.info('Image info: %s', image)
# Run detection
results = model.detect([image], verbose=1)
r = results[0]
logging.info('Runing model.detect([image], verbose=1).')
# Just apply mask then save images
print_img = visualize.apply_mask_instances(image, r['rois'], r['masks'],
r['class_ids'], class_names,
r['scores'], None, None, None,
None, None, [(1.0, 1.0, 1.0)])
skimage.io.imsave(saved_file_name, print_img)
logging.info('Finished apply_mask_instances.')
return True
def run_detect(filename):
base_file_name = os.path.basename(filename)
base_dir_name = os.path.dirname(filename)
logging.info('Dir name: %s',base_dir_name)
logging.info('File name: %s',base_file_name)
split_file_name, split_file_ext = os.path.splitext(base_file_name)
saved_dir_name = 'masked'
saved_file_name = str.join('\\', (base_dir_name, saved_dir_name, base_file_name))
logging.info('Saved as: %s',saved_file_name)
# Convert png with alpha channel with shape[2] == 4 into shape[2] ==3 RGB images
image = skimage.io.imread(filename)
if len(image.shape) > 2 and image.shape[2] == 4:
image = cv2.cvtColor(image, cv2.COLOR_BGRA2BGR)
# Run detection
results = model.detect([image], verbose=1)
r= results[0]
# Just apply mask then save images
print_img = visualize.apply_mask_instances(image, r['rois'], r['masks'], r['class_ids'], class_names, r['scores'])
skimage.io.imsave(saved_file_name,print_img)
return True
# convert png with alpha channel with shape[2] == 4 into shape[2] ==3 RGB images
image = skimage.io.imread(filename)
if len(image.shape) > 2 and image.shape[2] == 4:
image = cv2.cvtColor(image, cv2.COLOR_BGRA2BGR)
# processing time log
t0 = time.perf_counter()
# Run detection
results = model.detect([image], verbose=1)
r = results[0]
# without displaying images for batch program
###print_img = visualize.display_instances(image, r['rois'], r['masks'], r['class_ids'], class_names, r['scores'])
###skimage.io.imsave(saved_file_name,print_img)
# just apply mask then save images
print_img = visualize.apply_mask_instances(image, r['rois'],
r['masks'],
r['class_ids'],
class_names,
r['scores'])
skimage.io.imsave(saved_file_name, print_img)
# processing time log
t1 = time.perf_counter()
print('LOGGING:::::: Completed in %f sec' % (t1 - t0))
print('LOGGING:::::: End\n')
# end
def apply_mask_to_image(config_name, weights_name):
# Class names
# Index of the class in the list is its ID. For example, to get ID of
# the teddy bear class, use: class_names.index('teddy bear')
class_names = ['BG', 'numberplate']
# Create model object in inference mode.
model = modellib.MaskRCNN(mode="inference",
model_dir=MODEL_DIR,
config=config_name)
# Load pre-trained weights
NUMBERPLATE_WEIGHTS_FILE = os.path.join(NUMBERPLATE_WEIGHTS_PATH,
weights_name)
model.load_weights(NUMBERPLATE_WEIGHTS_FILE, by_name=True)
logging.info('%s weights file loaded.' % args.weights)
# Validation images extension
image_type_ok_list = ['jpeg', 'png', 'gif', 'bmp']
# Loop every file in the folder
file_count = 0
for file_names in os.scandir(IMAGE_DIR):
if file_names.is_file():
image_type = imghdr.what(file_names)
if image_type in image_type_ok_list:
# Start
file_count += 1
logging.info('Counting %d start' % file_count)
# Date time log
img_timestamp = time.localtime()
date_string = str.join('/',
(str(img_timestamp.tm_year),
str(img_timestamp.tm_mon).zfill(2),
str(img_timestamp.tm_mday).zfill(2)))
time_string = str.join(':',
(str(img_timestamp.tm_hour).zfill(2),
str(img_timestamp.tm_min).zfill(2),
str(img_timestamp.tm_sec).zfill(2)))
logging.info('Date Time: %s %s', date_string, time_string)
# Filename and filepath log
filename = os.path.join(IMAGE_DIR, file_names)
logging.info('Loading image: %s', filename)
# Convert png with alpha channel with shape[2] == 4 into shape[2] ==3 RGB images
image = skimage.io.imread(filename)
if len(image.shape) > 2 and image.shape[2] == 4:
image = cv2.cvtColor(image, cv2.COLOR_BGRA2BGR)
# Processing start time
t0 = time.perf_counter()
# Run detection
results = model.detect([image], verbose=1)
r = results[0]
# Just apply mask then save images
base_file_name = os.path.basename(filename)
base_dir_name = os.path.dirname(filename)
saved_dir_name = 'masked'
saved_file_name = str.join(
'\\', (base_dir_name, saved_dir_name, base_file_name))
print_img = visualize.apply_mask_instances(
image, r['rois'], r['masks'], r['class_ids'], class_names,
r['scores'])
skimage.io.imsave(saved_file_name, print_img)
logging.info('Saved masked image as: %s', saved_file_name)
# Processing ending time
t1 = time.perf_counter()
logging.info('Completed 1 image in %f sec', (t1 - t0))
logging.info('End')
return True