Ankur Goyal commited on
Commit
fcfd908
1 Parent(s): 27d0a44

Improve bounding box for highlight

Browse files
Files changed (1) hide show
  1. app.py +9 -10
app.py CHANGED
@@ -49,25 +49,24 @@ def lift_word_boxes(document):
49
  return document.context["image"][0][1]
50
 
51
 
52
- def expand_bbox(word_boxes, padding=0.1):
53
  if len(word_boxes) == 0:
54
  return None
55
 
56
  min_x, min_y, max_x, max_y = zip(*[x[1] for x in word_boxes])
57
  min_x, min_y, max_x, max_y = [min(min_x), min(min_y), max(max_x), max(max_y)]
58
- if padding != 0:
59
- padding = max((max_x - min_x) * padding, (max_y - min_y) * padding)
60
- min_x = max(0, min_x - padding)
61
- min_y = max(0, min_y - padding)
62
- max_x = max_x + padding
63
- max_y = max_y + padding
64
  return [min_x, min_y, max_x, max_y]
65
 
66
 
67
  # LayoutLM boxes are normalized to 0, 1000
68
- def normalize_bbox(box, width, height):
69
- pct = [c / 1000 for c in box]
70
- return [pct[0] * width, pct[1] * height, pct[2] * width, pct[3] * height]
 
 
 
 
 
71
 
72
 
73
  examples = [
 
49
  return document.context["image"][0][1]
50
 
51
 
52
+ def expand_bbox(word_boxes):
53
  if len(word_boxes) == 0:
54
  return None
55
 
56
  min_x, min_y, max_x, max_y = zip(*[x[1] for x in word_boxes])
57
  min_x, min_y, max_x, max_y = [min(min_x), min(min_y), max(max_x), max(max_y)]
 
 
 
 
 
 
58
  return [min_x, min_y, max_x, max_y]
59
 
60
 
61
  # LayoutLM boxes are normalized to 0, 1000
62
+ def normalize_bbox(box, width, height, padding=0.005):
63
+ min_x, min_y, max_x, max_y = [c / 1000 for c in box]
64
+ if padding != 0:
65
+ min_x = max(0, min_x - padding)
66
+ min_y = max(0, min_y - padding)
67
+ max_x = min(max_x + padding, 1)
68
+ max_y = min(max_y + padding, 1)
69
+ return [min_x * width, min_y * height, max_x * width, max_y * height]
70
 
71
 
72
  examples = [