yale_pca测试.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import numpy.linalg as linalg\n",
    "import matplotlib.pyplot as plt\n",
    "import cv2\n",
    "import random\n",
    "from PIL import Image\n",
    "import time\n",
    "%matplotlib inline\n",
    "start = 0\n",
    "end = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pca(date_mat, max_rank=200):\n",
    "    date_mat = np.float32(np.mat(date_mat))\n",
    "    mean_value = np.mean(date_mat, axis=0)\n",
    "    mean_removed = date_mat - mean_value\n",
    "    # cov_mat = mean_removed * mean_removed.T\n",
    "    cov_mat = np.cov(mean_removed, rowvar=1)\n",
    "    eig_vals, eig_vects = np.linalg.eig(cov_mat)\n",
    "    sort_vals = np.argsort(eig_vals)\n",
    "    select_vals = sort_vals[: -(max_rank + 1): -1]\n",
    "    select_vects = eig_vects[:, select_vals]\n",
    "    select_vects = mean_removed.T * select_vects\n",
    "    lowD = mean_removed * select_vects\n",
    "    return lowD, select_vects, mean_value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def knn(inX, dataSet, labels, k):\n",
    "    global start\n",
    "    start= time.time()\n",
    "    inX = np.array(inX)\n",
    "    dataSet = np.array(dataSet)\n",
    "    labels = np.array(labels)\n",
    "\n",
    "    dataSetSize = dataSet.shape[0]\n",
    "    diffMat = np.tile(inX, (dataSetSize, 1)) - dataSet\n",
    "    sqDiffMat = diffMat ** 2\n",
    "    sqDistance = sqDiffMat.sum(axis=1)\n",
    "    distance = sqDistance ** 0.5\n",
    "    sorteedDisttTndices = distance.argsort()\n",
    "    classCount = {}\n",
    "    for i in range(k):\n",
    "        voteIlabel = labels[sorteedDisttTndices[i]]\n",
    "        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1\n",
    "    sortedClassCount = sorted(\n",
    "        classCount.items(), key=lambda d: d[1], reverse=True)\n",
    "    global end\n",
    "    end= time.time()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loadImage(image_path):\n",
    "    # print(image_path)\n",
    "    image = np.array(Image.open(image_path))\n",
    "    # print(image.shape)\n",
    "    image = image.flatten()\n",
    "    # print(image.shape)\n",
    "    return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def loadData(test_image_path, data_path=\"yalefaces/yalefaces/\"):\n",
    "    train_data = []\n",
    "    train_lable = []\n",
    "    test_data = []\n",
    "    face_label = [\"centerlight\", \"glasses\", \"happy\", \"leftlight\", \"noglasses\", \"normal\",\n",
    "                  \"rightlight\", \"sad\", \"sleepy\", \"surprised\", \"wink\"]\n",
    "    test_data.append(loadImage(test_image_path))\n",
    "    for i in range(1, 16):\n",
    "        train_path = data_path + \"subject\" \n",
    "        if i < 10:\n",
    "            train_path += '0'\n",
    "        train_path += str(i)\n",
    "        for label in face_label:\n",
    "            train_image_path = train_path + '.'+ label + \".gif\"\n",
    "        if test_image_path == train_image_path:\n",
    "            continue\n",
    "        train_data.append(loadImage(train_image_path))\n",
    "        train_lable.append(train_path)\n",
    "    return train_data, train_lable, test_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def main(rank):\n",
    "    face_label = [\"centerlight\", \"glasses\", \"happy\", \"leftlight\", \"noglasses\", \"normal\",\n",
    "                  \"rightlight\", \"sad\", \"sleepy\", \"surprised\", \"wink\"]\n",
    "    person = random.randint(1, 15)\n",
    "    picture = random.randint(0, 10)\n",
    "    test_face = \"yalefaces/yalefaces/subject\" \n",
    "    if person < 10:\n",
    "        test_face += '0'\n",
    "    test_lable = test_face + str(person) \n",
    "    test_face = test_lable + \".\" + face_label[picture] + \".gif\"\n",
    "\n",
    "    train_data, train_lable, test_data = loadData(test_face)\n",
    "    lowD, select_vects, mean_value = pca(train_data, max_rank=rank)\n",
    "    test_data -= mean_value\n",
    "    test_data = np.mat(test_data) * np.mat(select_vects)\n",
    "    \n",
    "    best_match = knn(test_data, lowD, train_lable, 1)\n",
    "    \n",
    "    accuracy = [0,0]\n",
    "    print(best_match)\n",
    "    print(test_lable)\n",
    "    if best_match == test_lable:\n",
    "        accuracy[0] += 1\n",
    "    else:\n",
    "        accuracy[1] += 1\n",
    "        \n",
    "    return accuracy[0] / (accuracy[0] + accuracy[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n",
      "yalefaces/yalefaces/subject06\n",
      "None\n",
      "yalefaces/yalefaces/subject08\n",
      "None\n",
      "yalefaces/yalefaces/subject13\n",
      "None\n",
      "yalefaces/yalefaces/subject09\n",
      "None\n",
      "yalefaces/yalefaces/subject01\n",
      "None\n",
      "yalefaces/yalefaces/subject12\n",
      "None\n",
      "yalefaces/yalefaces/subject14\n",
      "None\n",
      "yalefaces/yalefaces/subject08\n",
      "None\n",
      "yalefaces/yalefaces/subject01\n",
      "None\n",
      "yalefaces/yalefaces/subject14\n",
      "None\n",
      "yalefaces/yalefaces/subject05\n",
      "None\n",
      "yalefaces/yalefaces/subject11\n",
      "None\n",
      "yalefaces/yalefaces/subject11\n",
      "None\n",
      "yalefaces/yalefaces/subject10\n",
      "None\n",
      "yalefaces/yalefaces/subject11\n",
      "None\n",
      "yalefaces/yalefaces/subject03\n",
      "None\n",
      "yalefaces/yalefaces/subject06\n",
      "None\n",
      "yalefaces/yalefaces/subject15\n",
      "None\n",
      "yalefaces/yalefaces/subject05\n",
      "None\n",
      "yalefaces/yalefaces/subject11\n",
      "None\n",
      "yalefaces/yalefaces/subject14\n",
      "None\n",
      "yalefaces/yalefaces/subject10\n",
      "None\n",
      "yalefaces/yalefaces/subject15\n",
      "None\n",
      "yalefaces/yalefaces/subject02\n",
      "None\n",
      "yalefaces/yalefaces/subject11\n",
      "None\n",
      "yalefaces/yalefaces/subject06\n",
      "None\n",
      "yalefaces/yalefaces/subject10\n",
      "None\n",
      "yalefaces/yalefaces/subject02\n",
      "None\n",
      "yalefaces/yalefaces/subject08\n",
      "None\n",
      "yalefaces/yalefaces/subject04\n",
      "[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZEAAAD8CAYAAAC2PJlnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f39400c55c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(1)\n",
    "t = []\n",
    "m = []\n",
    "rate, ends, starts = 0, 0, 0\n",
    "\n",
    "accuracy = []\n",
    "for i in range(1,100, 10):\n",
    "    rate, ends, starts = 0, 0, 0\n",
    "    for j in range(3):\n",
    "        rate += main(i)\n",
    "        ends += end\n",
    "        starts += start\n",
    "    accuracy.append(rate / 2)\n",
    "    t.append((ends - starts)/3)\n",
    "    m.append(i)\n",
    "    \n",
    "plt.subplot(121)  \n",
    "plt.plot(m,t, \".-\")\n",
    "plt.subplot(122) \n",
    "plt.plot(m,accuracy,\".-\")\n",
    "plt.draw()#注意此函数需要调用\n",
    "print(accuracy)\n"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "4ce0e62306dd6a5716965d4519ada776f947e6dfc145b604b11307c10277ef29"
  },
  "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}