face_read.py

import cv2
import pickle
import os
from utils import draw
import numpy as np

curr_path = os.getcwd()
Train_path = curr_path+'/WIDER_train/images'

H = 64
W = 64
cnt = 0

hog = cv2.HOGDescriptor((W, H), (16, 16), (8,8), (8,8), 9)
data = []
label = []

with open('./wider_face_split/wider_face_train_bbx_gt.txt', 'r') as f:
    curr_img = ''
    img = np.zeros((H, W))
    for line in f:
        inp = line.split(' ')
        if len(inp)==1:
            img_path = inp[0]
            img_path = img_path[:-1]
            curr_img = img_path
            if curr_img.isdigit():
                continue
            # print(Train_path+'/'+curr_img)
            img = cv2.imread(Train_path+'/'+curr_img, 2)
        else:
            inp = [int(i) for i in inp[:-1]]
            x1, y1, w, h, blur, expression, illumination, invalid, occlusion, pose = inp
            n = max(w,h)
            if invalid == 1 or blur > 0 or n < 50:
                continue
            img2 = img[y1:y1+n, x1:x1+n]
            img3 = cv2.resize(img2, (H, W))
            vec = hog.compute(img3)
            data.append(vec.flatten())
            label.append(1)
            cnt += 1
            print("{}: {} {} {} {}".format(len(vec),x1, y1, w, h))

with open('pos_data', 'wb') as f:
    pickle.dump((data, label), f)
print(cnt)