def blur_detector(img_col, thresh=10, mask=False):
assert isinstance(img_col, numpy.ndarray), 'img_col must be a numpy array'
assert img_col.ndim == 3, 'img_col must be a color image ({0} dimensions currently)'.format(img_col.ndim)
args = scripts.gen_args()
args.thresh = thresh
if mask:
return FocusMask.blur_mask(img)
else:
return evaluate(img_col=img_col, args=args)
def blur_detector(img_col, thresh=10, mask=False):
assert isinstance(img_col, numpy.ndarray), 'img_col must be a numpy array'
assert img_col.ndim == 3, 'img_col must be a color image ({0} dimensions currently)'.format(img_col.ndim)
args = scripts.gen_args()
args.thresh = thresh
if mask:
return FocusMask.blur_mask(img)
else:
return evaluate(img_col=img_col, args=args)
def process(image, save=False, display=False, args=None, segment=False):
assert isinstance(image, numpy.ndarray)
if not args:
args = scripts.gen_args()
else:
assert isinstance(args, argparse.Namespace), "args must be an argparse.Namespace"
args.save = save
args.display = display
detector = SkinDetector(args)
if segment:
slic = SpeedySuperPixels.SuperContour()
skin = numpy.zeros(image.shape, dtype=image.dtype)
for roi, contour in slic.process(image):
pxl = cv2.bitwise_and(image, image, mask=contour)
msk = detector.process(pxl)
ret = msk.sum() / contour.sum()
if ret > 0.8:
skin = numpy.min(255, cv2.add(skin, contour))
return skin
else:
return detector.process(image)
def process(image, save=False, display=False, args=None, segment=False):
assert isinstance(image, numpy.ndarray)
if not args:
args = scripts.gen_args()
else:
assert isinstance(args, argparse.Namespace), 'args must be an argparse.Namespace'
args.save = save
args.display = display
detector = SkinDetector(args)
if segment:
slic = SpeedySuperPixels.SuperContour()
skin = numpy.zeros(image.shape, dtype=image.dtype)
for roi, contour in slic.process(image):
pxl = cv2.bitwise_and(image, image, mask=contour)
msk = detector.process(pxl)
ret = msk.sum()/contour.sum()
if ret > 0.8:
skin = numpy.min(255, cv2.add(skin, contour))
return skin
else:
return detector.process(image)
