Skip to content

Commit

Permalink
remove old code, raster preds no longer saved
Browse files Browse the repository at this point in the history
  • Loading branch information
@Mary-h86
Mary-h86 committed Jul 29, 2023
1 parent e9d6877 commit 5c05043
Showing 1 changed file with 4 additions and 187 deletions.
191 changes: 4 additions & 187 deletions src/tile2net/tileseg/utils/misc.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -333,9 +333,9 @@ def dump(self, dump_dict, val_idx, testing=None, grid=None):
alpha = False
all_pix = (np.array(input_image).shape[0] * np.array(input_image).shape[1])
if np.array(input_image).shape[-1] == 3:
black = np.count_nonzero(np.all(np.array(input_image) == [50, 50, 50], axis=2))
black = np.count_nonzero(np.all(np.array(input_image) == [0, 0, 0], axis=2))
elif np.array(input_image).shape[-1] == 4:
black = np.count_nonzero(np.all(np.array(input_image) == [50, 50, 50, 50], axis=-1))
black = np.count_nonzero(np.all(np.array(input_image) == [0, 0, 0, 0], axis=-1))
alpha = True
ratio = black / all_pix

Expand All @@ -345,7 +345,7 @@ def dump(self, dump_dict, val_idx, testing=None, grid=None):
else:
prediction_pil = colorize_mask_fn(prediction)
prediction_pil = prediction_pil.convert('RGB')
self.create_composite_image(input_image, prediction_pil, img_name)
# self.create_composite_image(input_image, prediction_pil, img_name)

if grid:
idd_ = img_name.split('_')[-1]
Expand Down Expand Up @@ -375,190 +375,7 @@ def dump(self, dump_dict, val_idx, testing=None, grid=None):

self.imgs_to_tensorboard.append(to_tensorboard)

# def save_prob_and_err_mask(self, dump_dict, img_name, idx, prediction):
# if 'err_mask' in dump_dict and 'prob_mask' in dump_dict['assets']:
# prob_image = dump_dict['assets']['prob_mask'][idx]
# err_mask = dump_dict['err_mask'][idx]
# self.save_image((prob_image.cpu().numpy() * 255).astype(np.uint8), f'{img_name}_prob.png')
# err_pil = Image.fromarray(prediction.astype(np.uint8)).convert('RGB')
# err_pil.save(os.path.join(self.save_dir, f'{img_name}_err_mask.png'))
# return True
# return False
#
# def create_composite_image(self, input_image, prediction_pil, img_name):
# composited = Image.new('RGB', (input_image.width + input_image.width, input_image.height))
# composited.paste(input_image, (0, 0))
# composited.paste(prediction_pil, (prediction_pil.width, 0))
# composited_fn = 'sidebside_{}.png'.format(img_name)
# composited_fn = os.path.join(self.save_dir, composited_fn)
# composited.save(composited_fn)
#
# def dump_assets(self, dump_dict, img_name, idx):
# colorize_mask_fn = cfg.DATASET_INST.colorize_mask
# if self.dump_assets:
# assets = dump_dict['assets']
# for asset in assets:
# mask = assets[asset][idx]
# mask_fn = os.path.join(self.save_dir, f'{img_name}_{asset}.png')
# if 'pred_' in asset:
# pred_pil = colorize_mask_fn(mask)
# pred_pil.save(mask_fn)
# continue
# if type(mask) == torch.Tensor:
# mask = mask.squeeze().cpu().numpy()
# else:
# mask = mask.squeeze()
# mask = (mask * 255)
# mask = mask.astype(np.uint8)
# mask_pil = Image.fromarray(mask)
# mask_pil = mask_pil.convert('RGB')
# mask_pil.save(mask_fn)
# self.to_tensorboard.append(self.visualize(mask_pil))
#
# def dump(self, dump_dict, val_idx, testing=None, grid=None):
# """
# dump a single batch of images
#
# dump_dict: a dictionary containing elements to dump out
# 'input_images': source image
# 'gt_images': label
# 'img_names': img_names
# 'assets': dict with keys:
# 'predictions': final prediction
# 'pred_*': different scales of predictions
# 'attn_*': different scales of attn
# 'err_mask': err_mask
# """
# if (val_idx % self.dump_frequency or cfg.GLOBAL_RANK != 0):
# return
# else:
# pass
#
# colorize_mask_fn = cfg.DATASET_INST.colorize_mask
#
# for idx in range(len(dump_dict['input_images'])):
# input_image = dump_dict['input_images'][idx]
# #prob_image = dump_dict['assets']['prob_mask'][idx]
# gt_image = dump_dict['gt_images'][idx]
# prediction = dump_dict['assets']['predictions'][idx]
# #dump_dict['assets']['predictions']
# img_name = dump_dict['img_names'][idx]
#
# #for tile2net
# idd_ = img_name.split('_')[-1]
# er_prob = False
# if 'err_mask' in dump_dict and 'prob_mask' in dump_dict['assets']:
# err_mask = dump_dict['err_mask'][idx]
# prob_image = dump_dict['assets']['prob_mask'][idx]
#
# prob_fn = '{}_prob.png'.format(img_name)
# prob_fn = os.path.join(self.save_dir, prob_fn)
#
# cv2.imwrite(prob_fn, (prob_image.cpu().numpy()*255).astype(np.uint8))
#
# err_fn = f'{img_name}_err_mask.png'
# err_fn = os.path.join(self.save_dir,err_fn)
# err_pil = Image.fromarray(prediction.astype(np.uint8)).convert('RGB')
# err_pil.save(os.path.join(self.save_dir, err_fn))
# er_prob=True
#
#
# input_image = self.inv_normalize(input_image)
# input_image = input_image.cpu()
# input_image = standard_transforms.ToPILImage()(input_image)
# input_image = input_image.convert("RGB")
# # input_image_fn = f'{img_name}_input.png'
# # gt_fn = '{}_gt.png'.format(img_name)
# gt_pil = colorize_mask_fn(gt_image.cpu().numpy())
#
# if testing:
# alpha = False
# all_pix = (np.array(input_image).shape[0]*np.array(input_image).shape[1])
# if np.array(input_image).shape[-1] ==3 :
# black = np.count_nonzero(np.all(np.array(input_image) ==[50,50,50], axis=2))
# elif np.array(input_image).shape[-1] == 4:
# black = np.count_nonzero(np.all(np.array(input_image) == [50,50,50,50], axis=-1))
# alpha=True
# ratio = black/all_pix
#
# if ratio > 0.25 or alpha:
# continue
#
# else:
# prediction_pil = colorize_mask_fn(prediction)
# prediction_pil = prediction_pil.convert('RGB')
#
# composited = Image.new('RGB', (input_image.width + input_image.width, input_image.height))
# composited.paste(input_image, (0, 0))
# composited.paste(prediction_pil, (prediction_pil.width, 0))
# composited_fn = 'sidebside_{}.png'.format(img_name)
# if grid:
# # save_dir = grid.project.segmentation.path
# save_dir = os.path.join(cfg.RESULT_DIR, 'seg_results')
# self.save_dir = save_dir
# grid.tiles[grid.pose_dict[int(idd_)]].map_features(np.array(prediction_pil), img_array=True)
# #temporary
# composited_fn = os.path.join(self.save_dir, composited_fn)
# composited.save(composited_fn)
# else:
# composited_fn = os.path.join(self.save_dir, composited_fn)
# composited.save(composited_fn)
# # prediction_fn = '{}_pred.png'.format(img_name)
# else:
# gt_fn = '{}_gt.png'.format(img_name)
# gt_pil = colorize_mask_fn(gt_image.cpu().numpy())
# prediction_fn = '{}_prediction.png'.format(img_name)
# prediction_pil = colorize_mask_fn(prediction)
# prediction_pil = prediction_pil.convert('RGB')
# composited = Image.new('RGB', (input_image.width + input_image.width, input_image.height))
# composited.paste(input_image, (0, 0))
# composited.paste(prediction_pil, (prediction_pil.width, 0))
# composited_fn = 'sidebside_{}.png'.format(img_name)
# composited_fn = os.path.join(self.save_dir, composited_fn)
# composited.save(composited_fn)
#
#
# # only visualize a limited number of images
# if val_idx % self.viz_frequency or cfg.GLOBAL_RANK != 0:
# return
#
# to_tensorboard = [
# self.visualize(input_image.convert('RGB')),
# self.visualize(gt_pil.convert('RGB')),
# self.visualize(prediction_pil.convert('RGB')),
# ]
# if er_prob:
# to_tensorboard = [
# self.visualize(input_image.convert('RGB')),
# self.visualize(gt_pil.convert('RGB')),
# self.visualize(prediction_pil.convert('RGB')),
# self.visualize(err_pil.convert('RGB')),
# ]
#
#
# if self.dump_assets:
# assets = dump_dict['assets']
# for asset in assets:
# mask = assets[asset][idx]
# mask_fn = os.path.join(self.save_dir, f'{img_name}_{asset}.png')
#
# if 'pred_' in asset:
# pred_pil = colorize_mask_fn(mask)
# pred_pil.save(mask_fn)
# continue
#
# if type(mask) == torch.Tensor:
# mask = mask.squeeze().cpu().numpy()
# else:
# mask = mask.squeeze()
# mask = (mask * 255)
# mask = mask.astype(np.uint8)
# mask_pil = Image.fromarray(mask)
# mask_pil = mask_pil.convert('RGB')
# mask_pil.save(mask_fn)
# to_tensorboard.append(self.visualize(mask_pil))
#
# self.imgs_to_tensorboard.append(to_tensorboard)


def write_summaries(self, was_best):
"""
Expand Down

0 comments on commit 5c05043

Please sign in to comment.