def __getitem__(self, index):
img = ImageHelper.read_image(
self.img_list[index],
tool=self.configer.get('data', 'image_tool'),
mode=self.configer.get('data', 'input_mode'))
img_size = ImageHelper.get_size(img)
labelmap = ImageHelper.read_image(self.label_list[index],
tool=self.configer.get(
'data', 'image_tool'),
mode='P')
edgemap = ImageHelper.read_image(self.edge_list[index],
tool=self.configer.get(
'data', 'image_tool'),
mode='P')
edgemap[edgemap == 255] = 1
edgemap = cv2.resize(edgemap, (labelmap.shape[-1], labelmap.shape[-2]),
interpolation=cv2.INTER_NEAREST)
if self.configer.exists('data', 'label_list'):
labelmap = self._encode_label(labelmap)
if self.configer.exists('data',
'reduce_zero_label') and self.configer.get(
'data', 'reduce_zero_label') == 'True':
labelmap = self._reduce_zero_label(labelmap)
ori_target = ImageHelper.tonp(labelmap)
ori_target[ori_target == 255] = -1
if self.aug_transform is not None:
img, labelmap, edgemap = self.aug_transform(img,
labelmap=labelmap,
maskmap=edgemap)
border_size = ImageHelper.get_size(img)
if self.img_transform is not None:
img = self.img_transform(img)
if self.label_transform is not None:
labelmap = self.label_transform(labelmap)
edgemap = self.label_transform(edgemap)
meta = dict(ori_img_size=img_size,
border_size=border_size,
ori_target=ori_target)
return dict(
img=DataContainer(img, stack=True),
labelmap=DataContainer(labelmap, stack=True),
maskmap=DataContainer(edgemap, stack=True),
meta=DataContainer(meta, stack=False, cpu_only=True),
name=DataContainer(self.name_list[index],
stack=False,
cpu_only=True),
)
def __getitem__(self, index):
img = ImageHelper.read_image(
self.img_list[index],
tool=self.configer.get('data', 'image_tool'),
mode=self.configer.get('data', 'input_mode'))
img_size = ImageHelper.get_size(img)
labelmap = ImageHelper.read_image(self.label_list[index],
tool=self.configer.get(
'data', 'image_tool'),
mode='P')
if self.configer.exists('data', 'label_list'):
labelmap = self._encode_label(labelmap)
if self.configer.exists('data', 'reduce_zero_label'):
labelmap = self._reduce_zero_label(labelmap)
ori_target = ImageHelper.tonp(labelmap)
ori_target[ori_target == 255] = -1
ori_img = ImageHelper.tonp(img)
if self.aug_transform is not None:
img, labelmap = self.aug_transform(img, labelmap=labelmap)
border_size = ImageHelper.get_size(img)
if self.img_transform is not None:
img = self.img_transform(img)
if self.label_transform is not None:
labelmap = self.label_transform(labelmap)
meta = dict(ori_img_size=img_size,
border_size=border_size,
ori_target=ori_target,
ori_img=ori_img)
return dict(
img=DataContainer(img, stack=self.is_stack),
labelmap=DataContainer(labelmap, stack=self.is_stack),
meta=DataContainer(meta, stack=False, cpu_only=True),
name=DataContainer(self.name_list[index],
stack=False,
cpu_only=True),
)
def __getitem__(self, index):
img = ImageHelper.read_image(
self.img_list[index],
tool=self.configer.get('data', 'image_tool'),
mode=self.configer.get('data', 'input_mode'))
img_size = ImageHelper.get_size(img)
labelmap = ImageHelper.read_image(self.label_list[index],
tool=self.configer.get(
'data', 'image_tool'),
mode='P')
if self.configer.exists('data', 'label_list'):
labelmap = self._encode_label(labelmap)
distance_map, angle_map = self._load_maps(self.offset_list[index],
labelmap)
if self.configer.exists('data',
'reduce_zero_label') and self.configer.get(
'data', 'reduce_zero_label') == True:
labelmap = self._reduce_zero_label(labelmap)
ori_target = ImageHelper.tonp(labelmap).astype(np.int)
ori_target[ori_target == 255] = -1
ori_distance_map = np.array(distance_map)
ori_angle_map = np.array(angle_map)
if self.aug_transform is not None:
img, labelmap, distance_map, angle_map = self.aug_transform(
img,
labelmap=labelmap,
distance_map=distance_map,
angle_map=angle_map)
old_img = img
border_size = ImageHelper.get_size(img)
if self.img_transform is not None:
img = self.img_transform(img)
if self.label_transform is not None:
labelmap = self.label_transform(labelmap)
distance_map = torch.from_numpy(distance_map)
angle_map = torch.from_numpy(angle_map)
if set(self.configer.get('val_trans', 'trans_seq')) & set(
['random_crop', 'crop']):
ori_target = labelmap.numpy()
ori_distance_map = distance_map.numpy()
ori_angle_map = angle_map.numpy()
img_size = ori_target.shape[:2][::-1]
meta = dict(ori_img_size=img_size,
border_size=border_size,
ori_target=ori_target,
ori_distance_map=ori_distance_map,
ori_angle_map=ori_angle_map,
basename=os.path.basename(self.label_list[index]))
return dict(
img=DataContainer(img, stack=self.is_stack),
labelmap=DataContainer(labelmap, stack=self.is_stack),
distance_map=DataContainer(distance_map, stack=self.is_stack),
angle_map=DataContainer(angle_map, stack=self.is_stack),
meta=DataContainer(meta, stack=False, cpu_only=True),
name=DataContainer(self.name_list[index],
stack=False,
cpu_only=True),
)
def __getitem__(self, index):
img = ImageHelper.read_image(
self.img_list[index],
tool=self.configer.get('data', 'image_tool'),
mode=self.configer.get('data', 'input_mode'))
img_size = ImageHelper.get_size(img)
labelmap = ImageHelper.read_image(self.label_list[index],
tool=self.configer.get(
'data', 'image_tool'),
mode='P')
offsetmap_h = self._load_mat(self.offset_h_list[index])
offsetmap_w = self._load_mat(self.offset_w_list[index])
if os.environ.get('train_no_offset') and self.dataset == 'train':
offsetmap_h = np.zeros_like(offsetmap_h)
offsetmap_w = np.zeros_like(offsetmap_w)
if self.configer.exists('data', 'label_list'):
labelmap = self._encode_label(labelmap)
if self.configer.exists('data',
'reduce_zero_label') and self.configer.get(
'data', 'reduce_zero_label') == True:
labelmap = self._reduce_zero_label(labelmap)
# Log.info('use dataset {}'.format(self.configer.get('dataset')))
ori_target = ImageHelper.tonp(labelmap).astype(np.int)
ori_target[ori_target == 255] = -1
ori_offset_h = np.array(offsetmap_h)
ori_offset_w = np.array(offsetmap_w)
if self.aug_transform is not None:
img, labelmap, offsetmap_h, offsetmap_w = self.aug_transform(
img,
labelmap=labelmap,
offset_h_map=offsetmap_h,
offset_w_map=offsetmap_w)
border_size = ImageHelper.get_size(img)
if self.img_transform is not None:
img = self.img_transform(img)
if self.label_transform is not None:
labelmap = self.label_transform(labelmap)
offsetmap_h = torch.from_numpy(np.array(offsetmap_h)).long()
offsetmap_w = torch.from_numpy(np.array(offsetmap_w)).long()
meta = dict(
ori_img_size=img_size,
border_size=border_size,
ori_target=ori_target,
ori_offset_h=ori_offset_h,
ori_offset_w=ori_offset_w,
)
return dict(
img=DataContainer(img, stack=self.is_stack),
labelmap=DataContainer(labelmap, stack=self.is_stack),
offsetmap_h=DataContainer(offsetmap_h, stack=self.is_stack),
offsetmap_w=DataContainer(offsetmap_w, stack=self.is_stack),
meta=DataContainer(meta, stack=False, cpu_only=True),
name=DataContainer(self.name_list[index],
stack=False,
cpu_only=True),
)