-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathdraw_heat_new.py
173 lines (137 loc) · 4.32 KB
/
draw_heat_new.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
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
import numpy as np
from tqdm import tqdm
from sklearn.metrics.pairwise import cosine_similarity
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import os
from PIL import Image
from matplotlib.image import imread
# import cv2
import random
from nltk.tokenize import word_tokenize
from nltk import pos_tag
tag_dict = {}
tag_dict['CD'] = 1
tag_dict['JJ'] = 2
tag_dict['JJR'] = 2
tag_dict['JJS'] = 2
tag_dict['NN'] = 3
tag_dict['NNS'] = 3
tag_dict['NNP'] = 3
tag_dict['NNPS'] = 3
tag_dict['VB'] = 4
tag_dict['VBD'] = 4
tag_dict['VBG'] = 4
tag_dict['VBN'] = 4
tag_dict['VBP'] = 4
tag_dict['VBZ'] = 4
tag_dict['VBZ'] = 4
tag_dict['WP'] = 5
tag_dict['WP$'] = 5
tag_dict['WRB'] = 6
w = [1.41417003,0.65907812,1.98009944,1.50878048,0.68756366,1.72526932,1.37473035]
def sent_pos(sent):
# sent is a list of words
return [tag_dict[x[1]] if x[1] in tag_dict else 7 for x in pos_tag(sent)]
def nltk_tokenize(sent):
sent = str(sent).lower()
sent = sent.replace("-", " ")
sent = sent.replace("/", " ")
sent = sent.replace("`", " ")
token = word_tokenize(sent)
for i in range(len(token)):
if token[i].isalpha():
token[i] = token[i].lower()
return token
prefix = './'
file_list = sorted(os.listdir(prefix + 'att'))
# print file_list
xtmp = 0
# f, ax = plt.subplots(2,3,sharex=False,sharey=False)
# f, ax = plt.subplots(0,0,sharex=False,sharey=False)
def Generate_Heat(att):
x = np.zeros((224,224))
for idx in range(7):
for idy in range(7):
if att[idx][idy] > 0.7:
att[idx][idy] = 1
for idx in range(7):
for idy in range(7):
for i in range(224):
for j in range(224):
xx = idx * 32 + 16
yy = idy * 32 + 16
d = ((i - xx) ** 2 + (j - yy) ** 2) ** 0.5
x[i][j] += max(0,att[idx][idy] - d * 0.01)
return x
def Get_graph(ax, aword, gid):
xori = 1
for file in file_list:
qid,img,ques,ans = file.split('*')
qid = int(qid)
if qid != gid:
continue
graph_path = './graph/'+str(gid)
if os.path.exists(graph_path):
x = np.load(graph_path)
else:
att = np.load(prefix + 'att/' + file)
att = np.array(map(lambda x: ((x-x.min())/(x.max()-x.min())), att)).reshape(7,7)
x = Generate_Heat(att)
x.dump(graph_path)
xori = Image.open(prefix + 'images/' + img + '.jpg')
ori = xori.resize((224,224))
ax.imshow(ori)
ax.imshow(x, cmap=plt.cm.jet, alpha=0.9, interpolation='gaussian' )
ax.get_xaxis().set_visible(False)
ax.get_yaxis().set_visible(False)
ques = nltk_tokenize(str(ques).lower())[:-1]
ans = nltk_tokenize(str(ans).lower())[:-1]
ques_pos = sent_pos(ques)
ans_pos = sent_pos(ans)
quesx = np.arange(len(ques + ans))
quesy = [w[x - 1] for x in (ques_pos + ans_pos)]
print (ques + ans)
aword.plot(quesx, quesy,'ro-')
aword.set_xticks(quesx)
aword.set_xticklabels(ques + ans,rotation=30, fontsize = 10)
#aword.set_size_inches((3,3))
return xori
def Pic(ax, graph_id, i):
Get_graph(ax[1], ax[2], graph_id[1])
ori = Get_graph(ax[3], ax[4], graph_id[0])
ax[0].get_xaxis().set_visible(False)
ax[0].get_yaxis().set_visible(False)
if i==2 or i==4:
ax[0].plot(np.random.random(size=30))
ax[0].imshow( ori )
#graph_id =[[179456, 179457],[335448, 335449],[128800, 128801],[94908, 94909]]
good_id = [[100,101], [10052, 10053], [297766,297764], [32968, 32969], [202652, 202653]]#, [159696, 159697]] # [297740,297741],
bad_id = [[179456, 179457], [335448, 335449],[128800, 128801],[94908, 94909],[383988, 383989]]#, [324, 325]]
graph_id = good_id
# plt.tight_layout()
# plt.savefig(prefix + '2.eps',foramt = 'eps', dpi = 100, bbox_inches='tight')
from random import shuffle
file_list = os.listdir(prefix + 'att')
shuffle(file_list)
f, ax = plt.subplots(4,5, sharex=False,sharey=False)
plt.subplots_adjust(left=None, bottom=None, right=None, top=None,
wspace=-0.1, hspace=0.5)
f.set_size_inches(15.5, 10.5)
# for xid, file in enumerate(file_list[:2]):
# qid = int(file.split('*')[0])
# graph_id[xid] = [qid / 4 * 4, qid / 4 * 4 + 1]
print graph_id
i = 0
for tmp_ax, gid in zip(ax, graph_id):
Pic(tmp_ax, gid, i)
i=i+1
# plt.tight_layout()
plt.savefig(prefix + '2_4.eps',foramt = 'eps', dpi = 200, bbox_inches='tight')
#for num, gid in enumerate(graph_id):
# f, ax = plt.subplots(1,5,sharex=False,sharey=False)
# f.set_size_inches(18.5, 10.5)
# Pic(ax, gid)
# plt.tight_layout()
# plt.savefig(prefix + str(num) + '.eps',foramt = 'eps', dpi = 200)