wider_face数据格式说明
下载链接:http://shuoyang1213.me/WIDERFACE/index.html
说明:wider_face是从真实场景中选取出来的。其只标注出了人脸的矩形框区域,没有人脸对齐标注
code
import cv2
import os
xml_head = '''<annotation>
<folder>VOC2007</folder>
<!--文件名-->
<filename>{}</filename>.
<source>
<database>The VOC2007 Database</database>
<annotation>PASCAL VOC2007</annotation>
<image>flickr</image>
<flickrid>325991873</flickrid>
</source>
<owner>
<flickrid>null</flickrid>
<name>null</name>
</owner>
<size>
<width>{}</width>
<height>{}</height>
<depth>{}</depth>
</size>
<segmented>0</segmented>
'''
xml_obj = '''
<object>
<name>{}</name>
<pose>Rear</pose>
<!--是否被裁减,0表示完整,1表示不完整-->
<truncated>0</truncated>
<!--是否容易识别,0表示容易,1表示困难-->
<difficult>0</difficult>
<!--bounding box的四个坐标-->
<bndbox>
<xmin>{}</xmin>
<ymin>{}</ymin>
<xmax>{}</xmax>
<ymax>{}</ymax>
</bndbox>
</object>
'''
xml_end = '''
</annotation>'''
txt_path = '/home/yangna/data/WIDER_FACE/wider_face_split/wider_face_train_bbx_gt.txt'
imgs_path = '/home/yangna/data/WIDER_FACE/WIDER_train/images/'
labels = 'face' #label for datasets
cnt = 0
cnt_line = 0
with open(txt_path,'r') as train_list:
lst = train_list.readlines()
while (cnt_line < len(lst)):
img_name = lst[cnt_line]
if '.jpg' in img_name:
per_nums = int(lst[cnt_line + 1].rstrip())
per_lines = lst[cnt_line + 2:cnt_line+2+per_nums]
cnt_line = cnt_line + per_nums + 2
if per_nums == 0:
cnt_line += 1
continue
# process one image
img_path = imgs_path + img_name.rstrip()
img = cv2.imread(img_path)
img_h,img_w = img.shape[0],img.shape[1]
head = xml_head.format(str(img_path),str(img_w),str(img_h),3)
obj = ''
out_txt_name = './data/wider_face/voc/' + img_path.split('/')[-1].replace('.jpg', '.xml')
for f in per_lines:
f = [int(x) for x in f.rstrip().split(' ')]
if f[2]*f[3] < 10:
continue
xmin = str(f[0])
ymin = str(f[1])
xmax = str(f[0] + f[2])
ymax = str(f[1] + f[3])
obj += xml_obj.format(labels,xmin,ymin,xmax,ymax)
if obj == '':
continue
with open(out_txt_name,'w') as f_xml:
f_xml.write(head+obj+xml_end)
cnt += 1
print(cnt)
if cnt == 279:
temp = 1
result
处理后xml下载地址
百度云链接, 提取码: wrk3
wider_face带5个关键点的标注
下载链接:https://pan.baidu.com/s/1Laby0EctfuJGgGMgRRgykA
数据格式:
# 0--Parade/0_Parade_marchingband_1_849.jpg
449 330 122 149 488.906 373.643 0.0 542.089 376.442 0.0 515.031 412.83 0.0 485.174 425.893 0.0 538.357 431.491 0.0 0.82
# 0--Parade/0_Parade_Parade_0_904.jpg
361 98 263 339 424.143 251.656 0.0 547.134 232.571 0.0 494.121 325.875 0.0 453.83 368.286 0.0 561.978 342.839 0.0 0.89
# 0--Parade/0_Parade_marchingband_1_799.jpg
78 221 7 8 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 -1.0 0.2
格式 人脸矩形框(4个值),5个关键点(5×3个值),1个置信度值