Opencv EigenFace人臉識別算法詳解

2020-01-26 13:27:39

字體：大中小

來源：轉載

供稿：網友

簡要：

EigenFace是基于PCA降維的人臉識別算法，PCA是使整體數據降維后的方差最大，沒有考慮降維后類間的變化。它是將圖像每一個像素當作一維特征，然后用SVM或其它機器學習算法進行訓練。但這樣維數太多，根本無法計算。我這里用的是ORL人臉數據庫，英國劍橋實驗室拍攝的，有40位志愿者的人臉，在不同表情不同光照下每位志愿者拍攝10張，共有400張圖片，大小為112*92，所以如果把每個像素當做特征拿來訓練的話，一張人臉就有10304維特征，這么高維的數據根本無法處理。所以需要先對數據進行降維，去掉一些冗余的特征。

第一步：將ORL人臉圖片的地址統一放在一個文件里，等會通過對該文件操作，將圖片全部加載進來。

//ofstream一般對文件進行讀寫操作，ifstream一般對文件進行讀操作ofstream file; file.open("path.txt");//新建并打開文件 char str[50] = {}; for (int i = 1; i <= 40; i++) { for (int j = 1; j <= 10; j++) {   sprintf_s(str, "orl_faces/s%d/%d.pgm;%d", i, j, i);//將數字轉換成字符  file << str << endl;//寫入 }  }

得到路勁文件如下圖所示：

第二步：讀入模型需要輸入的數據，即用來訓練的圖像vector<Mat>images和標簽vector<int>labels

string filename = string("path.txt"); ifstream file(filename); if (!file) {     printf("could not load file");   } vector<Mat>images; vector<int>labels; char separator = ';'; string line,path, classlabel; while (getline(file,line)) { stringstream lines(line); getline(lines, path, separator); getline(lines, classlabel); images.push_back(imread(path, 0)); labels.push_back(atoi(classlabel.c_str()));//atoi(ASCLL to int)將字符串轉換為整數型 }

第三步：加載、訓練、預測模型

Ptr<BasicFaceRecognizer> model = EigenFaceRecognizer::create(); model->train(images, labels); int predictedLabel = model->predict(testSample); printf("actual label:%d,predict label :%d/n", testLabel, predictedLabel);

補充：

1、顯示平均臉

//計算特征值特征向量及平均值 Mat vals = model->getEigenValues();//89*1 printf("%d,%d/n", vals.rows, vals.cols); Mat vecs = model->getEigenVectors();//10324*89 printf("%d,%d/n", vecs.rows, vecs.cols); Mat mean = model->getMean();//1*10304 printf("%d,%d/n", mean.rows, mean.cols);  //顯示平均臉 Mat meanFace = mean.reshape(1, height);//第一個參數為通道數，第二個參數為多少行 normalize(meanFace, meanFace, 0, 255, NORM_MINMAX, CV_8UC1); imshow("Mean Face", meanFace);

2、顯示前部分特征臉

//顯示特征臉 for (int i = 0; i<min(10, vals.rows); i++) { Mat feature_vec = vecs.col(i).clone(); Mat feature_face= feature_vec.reshape(1, height);  normalize(feature_face, feature_face, 0, 255, NORM_MINMAX, CV_8UC1);  Mat colorface; applyColorMap(feature_face, colorface, COLORMAP_BONE);  sprintf_s(win_title, "eigenface%d", i); imshow(win_title, colorface); }

3、對第一張人臉在特征向量空間進行人臉重建（分別基于前10,20,30,40,50,60個特征向量進行人臉重建）

//重建人臉 for (int i = min(10, vals.rows); i <min(61, vals.rows); i+=10) { Mat vecs_space = Mat(vecs, Range::all(), Range(0, i)); Mat projection = LDA::subspaceProject(vecs_space, mean, images[0].reshape(1, 1));//投影到子空間 Mat reconstruction = LDA::subspaceReconstruct(vecs_space, mean, projection);//重建 Mat result = reconstruction.reshape(1, height); normalize(result, result, 0, 255, NORM_MINMAX, CV_8UC1); //char wintitle[40] = {}; sprintf_s(win_title, "recon face %d", i); imshow(win_title, result); }

完整代碼如下：

#include<opencv2/opencv.hpp>#include<opencv2/face.hpp>using namespace cv;using namespace face;using namespace std;char win_title[40] = {}; int main(int arc, char** argv) {  namedWindow("input",CV_WINDOW_AUTOSIZE);  //讀入模型需要輸入的數據，用來訓練的圖像vector<Mat>images和標簽vector<int>labels string filename = string("path.txt"); ifstream file(filename); if (!file) { printf("could not load file"); } vector<Mat>images; vector<int>labels; char separator = ';'; string line,path, classlabel; while (getline(file,line)) { stringstream lines(line); getline(lines, path, separator); getline(lines, classlabel); //printf("%d/n", atoi(classlabel.c_str())); images.push_back(imread(path, 0)); labels.push_back(atoi(classlabel.c_str()));//atoi(ASCLL to int)將字符串轉換為整數型 } int height = images[0].rows; int width = images[0].cols; printf("height:%d,width:%d/n", height, width); //將最后一個樣本作為測試樣本 Mat testSample = images[images.size() - 1]; int testLabel = labels[labels.size() - 1]; //刪除列表末尾的元素 images.pop_back(); labels.pop_back();  //加載，訓練，預測 Ptr<BasicFaceRecognizer> model = EigenFaceRecognizer::create(); model->train(images, labels); int predictedLabel = model->predict(testSample); printf("actual label:%d,predict label :%d/n", testLabel, predictedLabel);  //計算特征值特征向量及平均值 Mat vals = model->getEigenValues();//89*1 printf("%d,%d/n", vals.rows, vals.cols); Mat vecs = model->getEigenVectors();//10324*89 printf("%d,%d/n", vecs.rows, vecs.cols); Mat mean = model->getMean();//1*10304 printf("%d,%d/n", mean.rows, mean.cols);  //顯示平均臉 Mat meanFace = mean.reshape(1, height);//第一個參數為通道數，第二個參數為多少行 normalize(meanFace, meanFace, 0, 255, NORM_MINMAX, CV_8UC1); imshow("Mean Face", meanFace);  //顯示特征臉 for (int i = 0; i<min(10, vals.rows); i++) { Mat feature_vec = vecs.col(i).clone(); Mat feature_face= feature_vec.reshape(1, height);  normalize(feature_face, feature_face, 0, 255, NORM_MINMAX, CV_8UC1);  Mat colorface; applyColorMap(feature_face, colorface, COLORMAP_BONE);  sprintf_s(win_title, "eigenface%d", i); imshow(win_title, colorface); }  //重建人臉 for (int i = min(10, vals.rows); i <min(61, vals.rows); i+=10) { Mat vecs_space = Mat(vecs, Range::all(), Range(0, i)); Mat projection = LDA::subspaceProject(vecs_space, mean, images[0].reshape(1, 1)); Mat reconstruction = LDA::subspaceReconstruct(vecs_space, mean, projection); Mat result = reconstruction.reshape(1, height); normalize(result, result, 0, 255, NORM_MINMAX, CV_8UC1); //char wintitle[40] = {}; sprintf_s(win_title, "recon face %d", i); imshow(win_title, result); }  waitKey(0); return 0;}

以上就是本文的全部內容，希望對大家的學習有所幫助，也希望大家多多支持武林網。

上一篇：OpenCV實現輪廓的發現

下一篇：使用udp發送>=128K的消息會報ENOBUFS的錯誤的解決方法