-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.py
68 lines (57 loc) · 1.78 KB
/
utils.py
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
import itertools
import numpy as np
class Converter:
def __init__(self, alphabet):
pass
def labels_to_text(labels, alphabet):
"""Reverse translation of numerical classes back to characters."""
ret = []
for c in labels:
if c == 0: # CTC Blank
ret.append("")
else:
ret.append(alphabet[c - 1])
return "".join(ret)
def text_to_labels(text, alphabet):
"""Translation of characters to unique integer values"""
ret = []
for char in text:
ret.append(alphabet.find(char) + 1)
return ret
def decode_batch(test_func, word_batch, alphabet):
"""
- Greedy search -
For a real OCR application, this should be beam
search with a dictionary and language model.
For this example, best path is sufficient.
"""
res = []
prob = []
out = test_func([word_batch])[0]
for i in range(out.shape[0]):
out_best = list(np.argmax(out[i, 2:], 1))
out_prob = np.mean(np.max(out[i, 2:], 1))
out_best = [k for k, g in itertools.groupby(out_best)]
outstr = labels_to_text(out_best, alphabet)
res.append(outstr)
prob.append(out_prob)
return res, prob
def decode_batch2(out, alphabet):
"""
- Greedy search -
For a real OCR application, this should be beam
search with a dictionary and language model.
For this example, best path is sufficient.
"""
res = []
prob = []
print(out.size())
for i in range(out.shape[1]):
out_best = list(np.argmax(out[:, i], 1))
# out_prob = np.mean(np.max(out[:, i], 1))
out_prob = 1
out_best = [k for k, g in itertools.groupby(out_best)]
outstr = labels_to_text(out_best, alphabet)
res.append(outstr)
prob.append(out_prob)
return res, prob