gen_ALFW_data.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import time\n",
    "import numpy as np\n",
    "from numpy.random import randint\n",
    "import random\n",
    "\n",
    "def gen_pic(path,boxes,j,k):\n",
    "        \n",
    "    img=cv2.imread(path)\n",
    "    if img is None:\n",
    "        return j,k\n",
    "    h1,w1,_=img.shape\n",
    "    if(min(w1,h1)<=210):\n",
    "        return j,k\n",
    "    \n",
    "    num=9\n",
    "    while(num):\n",
    "        size=randint(100,min(w1,h1)/2)\n",
    "        x1=randint(1,w1-size)\n",
    "        y1=randint(1,h1-size)\n",
    "        box=np.array([x1,y1,x1+size,y1+size])\n",
    "        _boxes=boxes.copy()\n",
    "        if(np.max(IoU(box,np.array(_boxes)))<0.3):\n",
    "            resize_img=cv2.resize(img[y1:y1+size,x1:x1+size,:],(224,224))\n",
    "            cv2.imwrite('E:\\\\friedhelm\\\\Data\\\\non-face\\\\non_face_%d.jpg'%(j),resize_img)            \n",
    "            j=j+1\n",
    "            num=num-1\n",
    "\n",
    "    for bbox in boxes:\n",
    "        x=max(bbox[0],0)\n",
    "        y=max(bbox[1],0)\n",
    "        w=max(bbox[2],0)\n",
    "        h=max(bbox[3],0)\n",
    "        num=2 \n",
    "        while(num):\n",
    "\n",
    "            size=randint(np.floor(0.5*min(w,h))+1,np.ceil(1.5*max(w,h))+2)\n",
    "            x1=randint(np.floor(0.5*x)+1,np.ceil(1.5*x)+2)\n",
    "            y1=randint(np.floor(0.5*y)+1,np.ceil(1.5*y)+2)\n",
    "            box=np.array([x1,y1,x1+size,y1+size])\n",
    "\n",
    "            _bbox=np.array(bbox).reshape(1,-1)\n",
    "            if(IoU(box,_bbox)>0.65):\n",
    "                resize_img=cv2.resize(img[y1:y1+size,x1:x1+size,:],(224,224))\n",
    "                cv2.imwrite('E:\\\\friedhelm\\\\Data\\\\face\\\\face_%d.jpg'%(k),resize_img)            \n",
    "                k=k+1\n",
    "\n",
    "                if(random.choice([0,1])):\n",
    "                    resize_img=cv2.flip(resize_img,1)\n",
    "                    cv2.imwrite('E:\\\\friedhelm\\\\Data\\\\face\\\\face_%d.jpg'%(k),resize_img)             \n",
    "                    k=k+1\n",
    "                num=num-1  \n",
    "    return j,k\n",
    "\n",
    "def IoU(box,boxes):\n",
    "    \n",
    "    \"\"\"\n",
    "    box为实际，boxes为人脸\n",
    "    \"\"\"\n",
    "    area=boxes[:,2]*boxes[:,3]\n",
    "    \n",
    "    x1=np.maximum(box[0],boxes[:,0])\n",
    "    y1=np.maximum(box[1],boxes[:,1])    \n",
    "    x2=np.minimum(box[2],boxes[:,2]+boxes[:,0])\n",
    "    y2=np.minimum(box[3],boxes[:,3]+boxes[:,1]) \n",
    "    \n",
    "    w=np.maximum(0,x2-x1+1)\n",
    "    h=np.maximum(0,y2-y1+1)\n",
    "    return w*h/area\n",
    "\n",
    "def main():\n",
    "    begin=time.time()\n",
    "    addr='E:\\\\friedhelm\\\\AFLW\\\\'\n",
    "    with open(r'E:\\friedhelm\\alfw.txt') as f:\n",
    "        j=0\n",
    "        k=0\n",
    "        boxes=[]\n",
    "        path_compare=\"1\"\n",
    "        for line in f.readlines():\n",
    "            line=line.strip().split()\n",
    "            path=addr+line[0]\n",
    "            x=int(line[1])\n",
    "            if x<0:x=0\n",
    "            y=int(line[2])\n",
    "            if y<0:y=0\n",
    "            w=int(line[3])\n",
    "            h=int(line[4])\n",
    "\n",
    "            if(path_compare==path):\n",
    "                boxes.append([x,y,w,h])\n",
    "            else:\n",
    "                if(path_compare!=\"1\"):\n",
    "                    j,k=gen_pic(path_compare,boxes,j,k)\n",
    "                boxes=[]\n",
    "                path_compare=path\n",
    "                boxes.append([x,y,w,h])\n",
    "\n",
    "    print(time.time()-begin) \n",
    "    \n",
    "if __name__=='__main__':\n",
    "    main()"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}