Chris Oswald commited on
Commit
da94d71
·
1 Parent(s): b4f24a3

converted to sequence of images

Browse files
Files changed (1) hide show
  1. SPIDER.py +25 -5
SPIDER.py CHANGED
@@ -66,11 +66,11 @@ def subset_file_list(all_files: List[str], subset_ids: Set[int]):
66
 
67
  def standardize_3D_image(
68
  image: np.ndarray,
69
- resize_shape: Tuple[int, int, int],
70
  ) -> np.ndarray:
71
  """Aligns dimensions of image to be (height, width, channels); resizes
72
- images to values specified in resize_shape; and rescales pixel values
73
- to Uint8."""
74
  # Align height, width, channel dims
75
  if image.shape[0] < image.shape[2]:
76
  image = np.transpose(image, axes=[1, 2, 0])
@@ -519,6 +519,16 @@ class SPIDER(datasets.GeneratorBasedBuilder):
519
  image_array_original,
520
  resize_shape,
521
  )
 
 
 
 
 
 
 
 
 
 
522
 
523
  # Load .mha mask file
524
  mask_path = os.path.join(paths_dict['masks'], 'masks', example)
@@ -537,6 +547,16 @@ class SPIDER(datasets.GeneratorBasedBuilder):
537
  resize_shape,
538
  )
539
 
 
 
 
 
 
 
 
 
 
 
540
  # Extract overview data corresponding to image
541
  image_overview = overview_dict[scan_id]
542
 
@@ -547,8 +567,8 @@ class SPIDER(datasets.GeneratorBasedBuilder):
547
  return_dict = {
548
  'patient_id':patient_id,
549
  'scan_type':scan_type,
550
- 'image':image_array_standardized,
551
- 'mask':mask_array_standardized,
552
  'image_path':image_path,
553
  'mask_path':mask_path,
554
  'metadata':image_overview,
 
66
 
67
  def standardize_3D_image(
68
  image: np.ndarray,
69
+ resize_shape: Tuple[int, int],
70
  ) -> np.ndarray:
71
  """Aligns dimensions of image to be (height, width, channels); resizes
72
+ images to height/width values specified in resize_shape; and rescales
73
+ pixel values to Uint8."""
74
  # Align height, width, channel dims
75
  if image.shape[0] < image.shape[2]:
76
  image = np.transpose(image, axes=[1, 2, 0])
 
519
  image_array_original,
520
  resize_shape,
521
  )
522
+
523
+ # Split image array into sequence of 2D images
524
+ split_len = image_array_standardized.shape[-1]
525
+ images_seq = [
526
+ np.squeeze(arr) for arr in np.split(
527
+ image_array_standardized,
528
+ split_len,
529
+ axis=-1,
530
+ )
531
+ ]
532
 
533
  # Load .mha mask file
534
  mask_path = os.path.join(paths_dict['masks'], 'masks', example)
 
547
  resize_shape,
548
  )
549
 
550
+ # Split mask array into sequence of 2D images
551
+ split_len = mask_array_standardized.shape[-1]
552
+ masks_seq = [
553
+ np.squeeze(arr) for arr in np.split(
554
+ mask_array_standardized,
555
+ split_len,
556
+ axis=-1,
557
+ )
558
+ ]
559
+
560
  # Extract overview data corresponding to image
561
  image_overview = overview_dict[scan_id]
562
 
 
567
  return_dict = {
568
  'patient_id':patient_id,
569
  'scan_type':scan_type,
570
+ 'image':images_seq,
571
+ 'mask':masks_seq,
572
  'image_path':image_path,
573
  'mask_path':mask_path,
574
  'metadata':image_overview,