opensource/PaddleOCR
3
0
Fork 0
mirror of https://github.com/PaddlePaddle/PaddleOCR.git synced 2026-10-25 10:17:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

add refer for some code

Browse Source
This commit is contained in:
WenmuZhou
2021-11-04 18:23:23 +08:00
parent 9b9b2d60cf
commit fe92c51c42
14 changed files with 151 additions and 169 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ppocr/data/imaug/iaa_augment.py
+4 -13
View File
@@ -1,16 +1,7 @@
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This code is refer from:
https://github.com/WenmuZhou/DBNet.pytorch/blob/master/data_loader/modules/iaa_augment.py
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
ppocr/data/imaug/make_border_map.py
+4
View File
@@ -1,4 +1,8 @@
# -*- coding:utf-8 -*-
"""
This code is refer from:
https://github.com/WenmuZhou/DBNet.pytorch/blob/master/data_loader/modules/make_border_map.py
"""
from __future__ import absolute_import
from __future__ import division
ppocr/data/imaug/make_pse_gt.py
+22 -8
View File
@@ -12,6 +12,7 @@ from shapely.geometry import Polygon
__all__ = ['MakePseGt']
class MakePseGt(object):
r'''
Making binary mask from detection data with ICDAR format.
@@ -38,16 +39,20 @@ class MakePseGt(object):
text_polys *= scale
gt_kernels = []
for i in range(1,self.kernel_num+1):
for i in range(1, self.kernel_num + 1):
# s1->sn, from big to small
rate = 1.0 - (1.0 - self.min_shrink_ratio) / (self.kernel_num - 1) * i
text_kernel, ignore_tags = self.generate_kernel(image.shape[0:2], rate, text_polys, ignore_tags)
rate = 1.0 - (1.0 - self.min_shrink_ratio) / (self.kernel_num - 1
) * i
text_kernel, ignore_tags = self.generate_kernel(
image.shape[0:2], rate, text_polys, ignore_tags)
gt_kernels.append(text_kernel)
training_mask = np.ones(image.shape[0:2], dtype='uint8')
for i in range(text_polys.shape[0]):
if ignore_tags[i]:
cv2.fillPoly(training_mask, text_polys[i].astype(np.int32)[np.newaxis, :, :], 0)
cv2.fillPoly(training_mask,
text_polys[i].astype(np.int32)[np.newaxis, :, :],
0)
gt_kernels = np.array(gt_kernels)
gt_kernels[gt_kernels > 0] = 1
@@ -59,16 +64,25 @@ class MakePseGt(object):
data['mask'] = training_mask.astype('float32')
return data
def generate_kernel(self, img_size, shrink_ratio, text_polys, ignore_tags=None):
def generate_kernel(self,
img_size,
shrink_ratio,
text_polys,
ignore_tags=None):
"""
Refer to part of the code:
https://github.com/open-mmlab/mmocr/blob/main/mmocr/datasets/pipelines/textdet_targets/base_textdet_targets.py
"""
h, w = img_size
text_kernel = np.zeros((h, w), dtype=np.float32)
for i, poly in enumerate(text_polys):
polygon = Polygon(poly)
distance = polygon.area * (1 - shrink_ratio * shrink_ratio) / (polygon.length + 1e-6)
distance = polygon.area * (1 - shrink_ratio * shrink_ratio) / (
polygon.length + 1e-6)
subject = [tuple(l) for l in poly]
pco = pyclipper.PyclipperOffset()
pco.AddPath(subject, pyclipper.JT_ROUND,
pyclipper.ET_CLOSEDPOLYGON)
pco.AddPath(subject, pyclipper.JT_ROUND, pyclipper.ET_CLOSEDPOLYGON)
shrinked = np.array(pco.Execute(-distance))
if len(shrinked) == 0 or shrinked.size == 0:
ppocr/data/imaug/make_shrink_map.py
+4
View File
@@ -1,4 +1,8 @@
# -*- coding:utf-8 -*-
"""
This code is refer from:
https://github.com/WenmuZhou/DBNet.pytorch/blob/master/data_loader/modules/make_shrink_map.py
"""
from __future__ import absolute_import
from __future__ import division
ppocr/data/imaug/random_crop_data.py
+4
View File
@@ -1,4 +1,8 @@
# -*- coding:utf-8 -*-
"""
This code is refer from:
https://github.com/WenmuZhou/DBNet.pytorch/blob/master/data_loader/modules/random_crop_data.py
"""
from __future__ import absolute_import
from __future__ import division
ppocr/data/imaug/text_image_aug/augment.py
+4 -13
View File
@@ -1,16 +1,7 @@
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This code is refer from:
https://github.com/RubanSeven/Text-Image-Augmentation-python/blob/master/augment.py
"""
import numpy as np
from .warp_mls import WarpMLS
ppocr/data/imaug/text_image_aug/warp_mls.py
+4 -13
View File
@@ -1,16 +1,7 @@
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This code is refer from:
https://github.com/RubanSeven/Text-Image-Augmentation-python/blob/master/warp_mls.py
"""
import numpy as np