近几个月的笑脸识别研究过程中踩了很多坑，担心记录在本地容易不小心给删了，记录一份放在网上
回顾学习之路，程序上从C++开始，到执着于Python,从实现简单的图像剪切到自己构建卷积神经网络，筚路蓝缕，我心依旧。
以下为关于笑脸识别的个人自学记录，不具备科学的严谨性，仅作参考。
【程序：批量尺寸修改】#用于修改尺寸from skimage import data_dir,io,transform,colorimport numpy as npdef convert_gray(f,**args):
rgb=io.imread(f) #依次读取rgb图片
dst=transform.resize(rgb,(256,256)) #将图片大小转换为256*256
return dst
ImagePath='/users/liuzuoli/facedata/asix/spider2/sp'# 保存路径str='/users/liuzuoli/facedata/asix/spider2/*.jpg'coll = io.ImageCollection(str,load_func=convert_gray)for i in range(len(coll)):
io.imsave(ImagePath+'/'+np.str(i)+'.jpg',coll[i]) #循环保存图片
【程序：截取人脸的函数】**def** saveFaces(image_name):
faces = detectFaces(image_name)
**if** faces: #将人脸保存在save_dir目录下。
#Image模块：Image.open获取图像句柄，crop剪切图像(剪切的区域就是detectFaces返回的坐标)，save保存。
save_dir = image_name.split('.')[0]+"_faces"
os.mkdir(save_dir)
count = 0
**for** (x1,y1,x2,y2) **in** faces:
file_name = os.path.join(save_dir,str(count)+".jpg")
Image.open(image_name).crop((x1,y1,x2,y2)).save(file_name)
count+=1
【程序-人脸画出68个点】import dlib #人脸识别的库dlibimport numpy as np #数据处理的库numpyimport cv2 #图像处理的库OpenCv# dlib预测器detector = dlib.get_frontal_face_detector()#PREDICTOR_PATH = "/Users/liuzuoli/PycharmProjects/68dlib/shape_predictor_68_face_landmarks.dat"
predictor =dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')#这里出现了一个报错path="/users/liuzuoli/test/"# cv2读取图像img=cv2.imread(path+"pic3.JPG")# 取灰度img_gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)# 人脸数rectsrects = detector(img_gray, 0)for i in range(len(rects)):
landmarks = np.matrix([[p.x, p.y] for p in predictor(img, rects[i]).parts()])for idx, point in enumerate(landmarks): # 68点的坐标
pos = (point[0, 0], point[0, 1]) # 利用cv2.circle给每个特征点画一个圈，共68个
cv2.circle(img, pos, 6, color=(0, 255, 0)) # 利用cv2.putText输出1-68,font后面的参数可以调整
font = cv2.FONT_HERSHEY_SIMPLEX
cv2.putText(img, str(idx+1), pos, font, 0.6, (0, 0, 255), 1, cv2.LINE_AA)
cv2.namedWindow("img", 2)
cv2.imshow("img", img)
cv2.waitKey(0)
【人脸检测-未完全实现2】#include #include #include
#include
#include
#include
#include
#include
#include
#include
static CvMemStorage* storage = 0; //创建一个内存存储器，来统一管理各种动态对象的内存
static CvHaarClassifierCascade* cascade = 0; //分类器void detect_and_draw( IplImage* image ); //检测人脸并标记const char* cascade_name = "/usr/local/Cellar/opencv/3.4.1_2/share/OpenCV/haarcascades/haarcascade_frontalface_alt_tree.xml"; //分类器名称
int main( int argc, char** argv )
{
CvCapture* capture = 0; //视频的结构体IplImage *frame, *frame_copy = 0; //读取每一帧的结构体cascade = (CvHaarClassifierCascade*)cvLoad( cascade_name, 0, 0, 0 ); //载入分类器 if( !cascade )
{
fprintf( stderr, "ERROR: Could not load classifier cascade\n" );
fprintf( stderr, "Usage: facedetect --cascade=\"\"
[filename|camera_index]\n" ); return -1;
}
storage = cvCreateMemStorage(0);
cvNamedWindow( "result", 1 ); //1表示autosize
//检测视频capture = cvCaptureFromCAM(-1); //调用摄像头if( capture )
{ for(;;)
{ if( !cvGrabFrame( capture )) //从摄像头或者视频文件中抓取帧
break;
frame = cvRetrieveFrame( capture ); //取回由函数cvGrabFrame抓取的图像
if( !frame ) break; if( !frame_copy ) //复制图像
frame_copy = cvCreateImage( cvSize(frame->width,frame->height),
IPL_DEPTH_8U, frame->nChannels );
if( frame->origin == IPL_ORIGIN_TL ) //图像顶点是否在顶-左
cvCopy( frame, frame_copy, 0 ); else
cvFlip( frame, frame_copy, 0 ); //翻转
IplImage *equ = cvCreateImage(cvGetSize(frame_copy),8, 1);
IplImage *gray = cvCreateImage(cvGetSize(frame_copy), 8, 1);
cvCvtColor(frame_copy, gray, CV_BGR2GRAY); //转灰度图
cvEqualizeHist(gray, equ); //直方图均衡化
//cvNamedWindow("yuantu");
//cvNamedWindow("equ");
//cvShowImage("yuantu",gray);
//cvShowImage("equ",equ);
detect_and_draw( frame_copy ); //人脸检测并标记
if( cvWaitKey(1) >= 0 ) break; //cvReleaseImage(&gray);
//cvReleaseImage(&equ);
}
cvReleaseImage( &frame_copy );
cvReleaseCapture( &capture );
}
cvWaitKey(-1); //检测图片的时候，等待显示cvDestroyWindow("result");return 0;
}
void detect_and_draw( IplImage* img )
{static CvScalar colors[] = //用8种颜色标记人脸{
{0,0,255},
{0,128,255},
{0,255,255},
{0,255,0},
{255,128,0},
{255,255,0},
{255,0,0},
{255,0,255}
};
double scale = 1.2; //缩放因子IplImage* gray = cvCreateImage( cvSize(img->width,img->height), 8, 1 );
IplImage* small_img = cvCreateImage( cvSize( cvRound (img->width/scale), //四舍五入
cvRound (img->height/scale)), 8, 1 );
int i;
cvCvtColor( img, gray, CV_BGR2GRAY ); //转灰度图cvResize( gray, small_img, CV_INTER_LINEAR ); //调整大小cvEqualizeHist( small_img, small_img ); //使灰度图象直方图均衡化cvClearMemStorage( storage ); //重置if( cascade )
{
double t = (double)cvGetTickCount(); //返回从操作系统启动所经过的毫秒数
CvSeq* faces = cvHaarDetectObjects( small_img, cascade, storage, 1.1, 2, 0/*CV_HAAR_DO_CANNY_PRUNING*/,
cvSize(30, 30) );
printf("face's total is %d\n",faces->total);
t = (double)cvGetTickCount() - t;
printf( "detection time = %gms\n", t/((double)cvGetTickFrequency()*1000.) ); //cvGetTickFrequency()返回每秒的计时周期数
for( i = 0; i total : 0); i++ )
{
CvRect* r = (CvRect*)cvGetSeqElem( faces, i );
CvRect tr(r->x,r->y,r->width,r->height); //用矩形框框出
cvRectangle(img, cvPoint(r->x * scale, r->y * scale), cvPoint( (r->x + r->width) * scale, (r->y + r->height) * scale), colors[i%8], 3, 8, 0);
//用原型框出
CvPoint center;
int radius;
center.x = cvRound((r->x + r->width*0.5)*scale);
center.y = cvRound((r->y + r->height*0.5)*scale);
radius = cvRound((r->width + r->height)*0.25*scale);
cvCircle( img, center, radius, colors[i%8], 3, 8, 0 );
//用ROI截取人脸区域
cvSetImageROI(small_img, tr); //用缩放后的图，设置源图像ROI
CvSize size1 = cvSize(r->width, r->height);
IplImage* roi_img = cvCreateImage(size1,small_img->depth,small_img->nChannels);
cvCopy(small_img,roi_img); //复制图像
cvResetImageROI(small_img); //源图像用完后，清空ROI
cvNamedWindow("picture", CV_WINDOW_AUTOSIZE);
cvShowImage("picture", roi_img); //cvReleaseImage( &roi_img );
//cvDestroyWindow("picture");
}
}
cvShowImage( "result", img );
cvReleaseImage( &gray );
cvReleaseImage( &small_img );
}
作者：刘必王A6
链接：https://www.jianshu.com/p/9430dcaa592b
434
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
