利用python将alfw转换成bing标注格式

1.训练一个检测器（bing），要将alfw的数据，生成以下格式的正样本标注格式。

alfw本来是sqlit3的数据库格式，之前生成了.txt的。

利用该.txt文件生成标注文件

import os
import cv2


data_base = './JPEGImages/'


face_file = open('face_rect.txt')


face_file.next()


if not os.path.exists('Annotations/flickr/0'):
    os.makedirs('Annotations/flickr/0')
    
if not os.path.exists('Annotations/flickr/2'):
    os.makedirs('Annotations/flickr/2')
    
if not os.path.exists('Annotations/flickr/3'):
    os.makedirs('Annotations/flickr/3')


for line in face_file:
    spline = line.split()
    imageid = spline[0]
    path = spline[1]
    xmin = spline[2]
    ymin = spline[3]
    width = spline[4]
    height = spline[5]
    xmax = int(xmin) + int(width)
    ymax = int(ymin) + int(height)
    
    filename = path.split('.')[0]
    
    img = cv2.imread(data_base + path)
    img_width = img.shape[1]
    img_height = img.shape[0]
    
    if int(xmin) <= 0:
        print path, 'xmin', xmin
        xmin = '1'
        
    if int(ymin) <= 0:
        print path, 'ymin', ymin
        ymin = '1'
        
    if xmax >= img_width:
        print path, 'xmax', xmax, 'img_width', img_width, 'img_height', img_height
        xmax = img_width-1
        
    if ymax >= img_height:
        print path, 'ymax', ymax
        ymax = img_height-1 
    
    if not os.path.exists('Annotations/'+filename+'.yml'):
        file = open('Annotations/'+filename+'.yml', 'w')
        print >> file, '%YAML:1.0\n'
        print >> file, 'annotation:'
        print >> file, '  folder: ALFW'
        print >> file, '  filename: \"%s\"'%path
        print >> file, '  source: {id: %s}' %imageid
        print >> file, '  owner: {name: zhuqian}'
        print >> file, '  size: {width: \'%s\', height: \'%s\', depth: \'3\'}'%(img_width, img_height)
        print >> file, '  segmented: \'0\''
        print >> file, '  object:'
        print >> file, '    - bndbox: {xmin: \'%s\', ymin: \'%s\', xmax: \'%s\', ymax: \'%s\'}'%(xmin, ymin, xmax, ymax)
        print >> file, '      name: face'
        print >> file, '      pose: Left'
        print >> file, '      truncated: \'1\''
        print >> file, '      difficult: \'0\''
        
    else:
        file = open('Annotations/'+filename+'.yml', 'a')
        print >> file, '    - bndbox: {xmin: \'%s\', ymin: \'%s\', xmax: \'%s\', ymax: \'%s\'}'%(xmin, ymin, xmax, ymax)
        print >> file, '      name: face'
        print >> file, '      pose: Left'
        print >> file, '      truncated: \'1\''
        print >> file, '      difficult: \'0\''
        
        #print path
        
    
    file.close()
    
face_file.close()

2. 随机生成训练测试样本。

import os
import random


images = []


with open('face_rect.txt','r') as face_file:
    face_file.next()   
    for line in face_file:
        img_name = line.split()[1].split('.')[0]
        images.append(img_name)


random.shuffle(images)


num = len(images)


with open('ImageSets/Main/train.txt','w') as train_file:
    for i in xrange(0, num/4):
        print >> train_file, images[i]
        
with open('ImageSets/Main/test.txt','w') as test_file:    
    for i in xrange(num/4+1, num):
        print >> test_file, images[i]
   
with open('ImageSets/Main/class.txt','w') as class_file:  
    print >> class_file, 'face'

3. 将不是jpg格式的图片转成jpg。

import os
import cv2

data_base = './JPEGImages/'

face_file = open('face_rect.txt')

face_file.next()

for line in face_file:
    spline = line.split()
    path = spline[1]
    
    filename = path.split('.')[0]
    ext = path.split('.')[1]
    
    if ext != 'jpg':
        print path
        img = cv2.imread(data_base + path)
        cv2.imwrite(data_base + filename + '.jpg', img)
    
face_file.close()