/
preprocess_detect_faces.py
96 lines (86 loc) · 3.41 KB
/
preprocess_detect_faces.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
"""
Disclaimer:
This code is based on code by Peter Ruch, see his prunhild repository
See: https://github.com/gfrogat/prunhild
Snippets of code also borrowed from Arun Joseph pruning code.
https://github.com/00arun00/Pruning-Pytorch/blob/master/prune.py
"""
import numpy as np
from mtcnn_pytorch_master.src import detect_faces, show_bboxes
from PIL import Image
import random
import os
def pp_save_clean_dataset():
directory = '/Users/guillaume/Documents/Soft/Project_Insight/Data/GOT'
dirname_ext = '_raw'
ftrain = open(directory+'/train.txt','w')
fval = open(directory+'/val.txt',"w")
ftest = open(directory+'/test.txt','w')
##file.write(“Hello World”)
#file.write(“This is our new text file”)
#file.write(“and this is another line.”)
#file.write(“Why? Because we can.”)
#file.close()
jclass=0
for filedir in os.listdir(directory):
if dirname_ext in filedir:
jclass+=1
i = 0
print(filedir)
for filename in os.listdir(directory+'/'+filedir):
#print(directory+'/'+filedir+'/'+filename)
if filename[0] != '.':
rnd = random.random()
# 70% train set, 20% val set, 10% test set.
if rnd<.7:
ftrain.write(directory+'/'+filedir+'/'+filename+' '+str(jclass)+'\n')
elif rnd<.9:
fval.write(directory+'/'+filedir+'/'+filename+' '+str(jclass)+'\n')
else:
ftest.write(directory+'/'+filedir+'/'+filename+' '+str(jclass)+'\n')
i+=1
ftrain.close()
fval.close()
ftest.close()
def pp_detect_faces():
hsize = 144
wsize = 144
directory = '/Users/guillaume/Documents/Soft/Project_Insight/Data/GOT'
dirname_ext = '_unprocessed'
for filedir in os.listdir(directory):
if dirname_ext in filedir:
i = 0
print(filedir)
for filename in os.listdir(directory+'/'+filedir):
#print(directory+'/'+filedir+'/'+filename)
if filename[0] != '.':
image = Image.open(directory+'/'+filedir+'/'+filename)
bounding_boxes, landmarks = detect_faces(image)
#print(len(bounding_boxes))
#print('---')
#print(bounding_boxes)
for ind in range(len(bounding_boxes)):
imgcopy = image.copy()
imgcopy = imgcopy.crop((bounding_boxes[ind,0], bounding_boxes[ind,1], bounding_boxes[ind,2], bounding_boxes[ind,3]))#.save(...)
imgcopy = imgcopy.resize((wsize,hsize))
imgcopy = imgcopy.convert("L")
outdir = filedir.replace(dirname_ext, '')
imgcopy.save(directory+'/'+outdir+'_raw'+'/'+str(i)+'.jpeg', "JPEG")
i+=1
#if i % 10 ==0:
#print(filename)
pp_detect_faces()
for filename in os.listdir(directory+'/'+filedir):
print(filename)
if i>10:
break
i+=1
###
# image = Image.open('./mtcnn_pytorch_master/images/example.png')
# bounding_boxes, landmarks = detect_faces(image)
# a = show_bboxes(image, bounding_boxes, landmarks)
# a.show()
###
#import prunhild
#from config import parser
#from utils import get_parameter_stats, print_parameter_stats