def __call__(self, img: np.ndarray, data: dict):
boxes = data['gt_boxes']
labels = data['gt_labels']
scores = data['gt_scores']
if len(boxes) == 0:
return img, data
if not self.constraints:
self.constraints = [(0.1, 1.0), (0.3, 1.0), (0.5, 1.0), (0.7, 1.0),
(0.9, 1.0), (0.0, 1.0)]
img = PIL.Image.fromarray(img)
w, h = img.size
crops = [(0, 0, w, h)]
for min_iou, max_iou in self.constraints:
for _ in range(self.max_trial):
scale = random.uniform(self.scales[0], self.scales[1])
aspect_ratio = random.uniform(max(1 / self.max_ratio, scale * scale), \
min(self.max_ratio, 1 / scale / scale))
crop_h = int(h * scale / np.sqrt(aspect_ratio))
crop_w = int(w * scale * np.sqrt(aspect_ratio))
crop_x = random.randrange(w - crop_w)
crop_y = random.randrange(h - crop_h)
crop_box = np.array([[(crop_x + crop_w / 2.0) / w,
(crop_y + crop_h / 2.0) / h,
crop_w / float(w), crop_h / float(h)]])
iou = box_iou_xywh(crop_box, boxes)
if min_iou <= iou.min() and max_iou >= iou.max():
crops.append((crop_x, crop_y, crop_w, crop_h))
break
while crops:
crop = crops.pop(np.random.randint(0, len(crops)))
crop_boxes, crop_labels, crop_scores, box_num = box_crop(
boxes, labels, scores, crop, (w, h))
if box_num < 1:
continue
img = img.crop((crop[0], crop[1], crop[0] + crop[2],
crop[1] + crop[3])).resize(img.size,
PIL.Image.LANCZOS)
img = np.asarray(img)
data['gt_boxes'] = crop_boxes
data['gt_labels'] = crop_labels
data['gt_scores'] = crop_scores
return img, data
img = np.asarray(img)
data['gt_boxes'] = boxes
data['gt_labels'] = labels
data['gt_scores'] = scores
return img, data
def crop_resize_image(image: np.ndarray, size) -> np.ndarray:
"""
Resize the input image.
:param image: Original image which is a PIL object.
:param size: Tuple of height and width to resize the image to.
:return: Resized image which is a PIL object
"""
width, height = image.size
if width > height:
left = (width - height) / 2
right = width - left
top = 0
bottom = height
else:
top = (height - width) / 2
bottom = height - top
left = 0
right = width
image = image.crop((left, top, right, bottom))
image = image.resize(size, Image.ANTIALIAS)
return image
def crop_resize_image(image: np.ndarray, size) -> np.ndarray:
"""
Resize the input image.
:param image: Original image which is a PIL object.
:param size: Tuple of height and width to resize the image to.
:return: Resized image which is a PIL object
"""
width, height = image.size
if width > height:
left = (width - height) / 2
right = width - left
top = 0
bottom = height
else:
top = (height - width) / 2
bottom = height - top
left = 0
right = width
image = image.crop((left, top, right, bottom))
image = image.resize(size, Image.ANTIALIAS)
return image
def crop(img: np.ndarray, x_pos: int, y_pos: int, size: int) -> np.ndarray:
ow, oh = img.size
if (ow > size or oh > size):
return img.crop((x_pos, y_pos, x_pos + size, y_pos + size))
return img
