/
collate.pyx
68 lines (58 loc) · 2.25 KB
/
collate.pyx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
from letters import Letter, TextLine
import numpy as np
cimport numpy as np
cimport cython
cdef class CLetter:
cdef np.ndarray label_map, stats, centroid
cdef int label, x, y, w, h, r, b
def __init__(self, letter):
self.label = letter.label
self.label_map = letter.label_map
self.stats = letter.stats
self.centroid = letter.centroid
self.x = letter.x
self.y = letter.y
self.w = letter.w
self.h = letter.h
self.r = self.x + self.w
self.b = self.y + self.h
def letter(self):
return Letter(self.label, self.label_map, self.stats, self.centroid)
def collate_lines(int AH, list letters):
cdef int score, best_score, line_len
cdef CLetter last1, last2, cl, letter, first, last
cdef list lines, best_candidate
letters = [CLetter(l) for l in letters]
letters.sort(key=lambda CLetter cl: cl.x)
lines = []
for letter in letters:
best_candidate = []
best_score = 100000
for line in lines:
line_len = len(line)
last1 = line[-1]
last2 = line[-2] if line_len > 1 else last1
score = best_score
if letter.x < last1.r + 4 * AH \
and last1.y <= letter.b + 3 and letter.y <= last1.b + 3:
score = letter.x - last1.r + abs(letter.y - last1.y)
elif line_len > 1 \
and letter.x < last2.x + last2.w + AH \
and last2.y <= letter.b and letter.y <= last2.b:
score = letter.x - last2.r + abs(letter.y - last2.y)
if score < best_score - 0.1:
best_score = score
best_candidate = line
if best_candidate:
first = best_candidate[0]
last = best_candidate[-1]
best_candidate.append(letter)
if letter.x - last.x > 300 or letter.x < first.x or letter.x - first.x > 2000:
print "agggghhh"
print "first", first.letter()
print "last ", last.letter()
print "new ", letter.letter()
# print " selected:", x, y, w, h
else:
lines.append([letter])
return [TextLine([cl.letter() for cl in line]) for line in lines]