SFaceRecognizer model face recognition C++ test failed

[Linktin]

YuNet model("/home/hy/code/face_detection_yunet_2023mar.onnx", cv::Size(160, 120), 0.9, 0.3, 5000, 0, 0);
SFace sface("/home/hy/code/face_recognition_sface_2021dec_int8.onnx",0,0,0);
DetectedFace dface1,dface2;

Mat img1 = imread("/home/hy/code/heye.jpg");
Mat img2 = imread("/home/hy/code/demo.jpg");

//face detection
model.setInputSize(Size(img1.cols,img1.rows));
dface1 = model.infer(img1);
model.setInputSize(Size(img2.cols,img2.rows));
dface2 = model.infer(img2);

std::cout<<"face1 nums:"<<dface1.nums<<std::endl;
std::cout<<"face2 nums:"<<dface2.nums<<std::endl;
//face recognition
auto tick_meter = cv::TickMeter();
tick_meter.start();

// std::cout<<"scores:"<<static_cast(dface1.faces.at(0, 14))<<std::endl;
cv::Range cols_range(0, 13);
cv::Mat face1_features = dface1.faces.row(0).colRange(cols_range);
cv::Mat face2_features = dface2.faces.row(0).colRange(cols_range);

pair<double, bool> res = sface.match(img1, face1_features, img2, face2_features);

When I use the image and the detected facial feature information using YuNet for inference with this model, the result is as follows:

In any case, the facial matching score is always 1! This is not normal.

[Linktin]

已解决，feature的浅拷贝导致得到相同的人脸特征，加上clone()函数进行深拷贝即可解决。
Mat feature1 = infer(image1,face1).clone(); // 提取第一张图像的特征
Mat feature2 = infer(image2,face2).clone(); // 提取第二张图像的特征

[fengyuentau]

Duplicate of #189