Skip to content

Commit

Permalink
Fix LetterResize bug with imrescale api (#105)
Browse files Browse the repository at this point in the history 
* [Feature] Make docker image smaller (#67)

* [Feature] Make docker image smaller


* Improve doc

* Use Compose directly in BaseMixImageTransform (#71)

* Use 'Compose' directly

Use 'Compose' directly

* avoids some unnecessary calculations

avoids some unnecessary calculations.

* remove whitespace

remove whitespace

* fix resize bug

* fix resize bug

* fix resize bug

* fix

* fix resize bug

* add test letterresize without batchshape

* fix resize bug

Co-authored-by: HinGwenWoong <peterhuang0323@qq.com>
Co-authored-by: jason_w <wongzheng@126.com>
  • Loading branch information
@Nioolek @PeterH0323
@MambaWong @hhaAndroid
3 people authored and hhaAndroid committed Sep 29, 2022
1 parent bb3aa48 commit 0d97ffe
Show file tree
Hide file tree
Showing 2 changed files with 57 additions and 22 deletions.
38 changes: 17 additions & 21 deletions mmyolo/datasets/transforms/transforms.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -167,13 +167,14 @@ def _resize_img(self, results: dict):

# Use batch_shape if a batch_shape policy is configured
if 'batch_shape' in results:
self.scale = tuple(results['batch_shape'])
scale = tuple(results['batch_shape'])
else:
scale = self.scale

image_shape = image.shape[:2] # height, width

# Scale ratio (new / old)
ratio = min(self.scale[0] / image_shape[0],
self.scale[1] / image_shape[1])
ratio = min(scale[0] / image_shape[0], scale[1] / image_shape[1])

# only scale down, do not scale up (for better test mAP)
if not self.allow_scale_up:
Expand All @@ -187,7 +188,7 @@ def _resize_img(self, results: dict):

# padding height & width
padding_h, padding_w = [
self.scale[0] - no_pad_shape[0], self.scale[1] - no_pad_shape[1]
scale[0] - no_pad_shape[0], scale[1] - no_pad_shape[1]
]
if self.use_mini_pad:
# minimum rectangle padding
Expand All @@ -196,20 +197,14 @@ def _resize_img(self, results: dict):
elif self.stretch_only:
# stretch to the specified size directly
padding_h, padding_w = 0.0, 0.0
no_pad_shape = (self.scale[0], self.scale[1])
ratio = [
self.scale[0] / image_shape[0], self.scale[1] / image_shape[1]
] # height, width ratios

# divide padding into 2 sides
padding_h /= 2
padding_w /= 2
no_pad_shape = (scale[0], scale[1])
ratio = [scale[0] / image_shape[0],
scale[1] / image_shape[1]] # height, width ratios

if image_shape[::-1] != no_pad_shape:
if image_shape != no_pad_shape:
# compare with no resize and padding size
image = mmcv.imrescale(
image,
no_pad_shape,
image = mmcv.imresize(
image, (no_pad_shape[1], no_pad_shape[0]),
interpolation=self.interpolation,
backend=self.backend)

Expand All @@ -221,10 +216,10 @@ def _resize_img(self, results: dict):
results['scale_factor'] = scale_factor

# padding
top_padding, bottom_padding = int(round(padding_h - 0.1)), int(
round(padding_h + 0.1))
left_padding, right_padding = int(round(padding_w - 0.1)), int(
round(padding_w + 0.1))
top_padding, left_padding = int(round(padding_h // 2 - 0.1)), int(
round(padding_w // 2 - 0.1))
bottom_padding = padding_h - top_padding
right_padding = padding_w - left_padding

padding_list = [
top_padding, bottom_padding, left_padding, right_padding
Expand Down Expand Up @@ -257,7 +252,8 @@ def _resize_masks(self, results: dict):
results['scale_factor'][0]
gt_mask_width = results['gt_masks'].width * \
results['scale_factor'][1]
gt_masks = results['gt_masks'].rescale((gt_mask_height, gt_mask_width))
gt_masks = results['gt_masks'].resize(
(int(round(gt_mask_height)), int(round(gt_mask_width))))

# padding the gt_masks
if len(gt_masks) == 0:
Expand Down
41 changes: 40 additions & 1 deletion tests/test_datasets/test_transforms/test_transforms.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -75,12 +75,51 @@ def test_letter_resize(self):
# Test stretch_only
transform = LetterResize(scale=(640, 640), stretch_only=True)
results = transform(copy.deepcopy(self.data_info1))
self.assertEqual(results['img_shape'], (460, 613, 3))
self.assertEqual(results['img_shape'], (460, 672, 3))
self.assertTrue((results['gt_bboxes'] == np.array(
[[0., 0., 230., 251.99998474121094]])).all())
self.assertTrue((results['batch_shape'] == np.array([460, 672])).all())
self.assertTrue((results['pad_param'] == np.array([0, 0, 0, 0])).all())

# Test
transform = LetterResize(scale=(640, 640), pad_val=dict(img=144))
rng = np.random.RandomState(0)
for _ in range(20):
input_h, input_w = np.random.randint(100, 700), np.random.randint(
100, 700)
output_h, output_w = np.random.randint(100,
700), np.random.randint(
100, 700)
data_info = dict(
img=np.random.random((input_h, input_w, 3)),
gt_bboxes=np.array([[0, 0, 10, 10]], dtype=np.float32),
batch_shape=np.array([output_h, output_w], dtype=np.int64),
gt_masks=BitmapMasks(
rng.rand(1, input_h, input_w),
height=input_h,
width=input_w))
results = transform(data_info)
self.assertEqual(results['img_shape'], (output_h, output_w, 3))
self.assertTrue(
(results['batch_shape'] == np.array([output_h,
output_w])).all())

# Test without batchshape
transform = LetterResize(scale=(640, 640), pad_val=dict(img=144))
rng = np.random.RandomState(0)
for _ in range(20):
input_h, input_w = np.random.randint(100, 700), np.random.randint(
100, 700)
data_info = dict(
img=np.random.random((input_h, input_w, 3)),
gt_bboxes=np.array([[0, 0, 10, 10]], dtype=np.float32),
gt_masks=BitmapMasks(
rng.rand(1, input_h, input_w),
height=input_h,
width=input_w))
results = transform(data_info)
self.assertEqual(results['img_shape'], (640, 640, 3))


class TestYOLOv5KeepRatioResize(unittest.TestCase):

Expand Down

0 comments on commit 0d97ffe

Please sign in to comment.