License Plates not getting detected successfully.

zhubenfu / License-Plate-Detect-Recognition-via-Deep-Neural-Networks-accuracy-up-to-99.9

works in real-time with detection and recognition accuracy up to 99.8% for Chinese license plates: 100 ms/plate

1.38k stars 325 forks source link

License Plates not getting detected successfully. #12

Open nanditam1 opened 5 years ago

nanditam1 commented 5 years ago

Hi, This is an interesting application of mtcnn!

I have tried running the caffemodels on a simple car image , however the plate detection is getting missed after passing 48-net stage. Also faces are getting detected! Are the following caffemodels for face detection? :- train_mtcnn_LPR / 12train_wm_lpr /det1.caffemodel train_mtcnn_LPR / 24train_wm_lpr /det2.caffemodel If so could you please share the license plate caffemodels? Also the regressed points are kind of random. Thank you!

zhubenfu commented 5 years ago

License-Plate-Detect-Recognition-via-Deep-Neural-Networks-accuracy-up-to-99.9/caffe-vsproj/ocr_test/model_platecar/

zhubenfu commented 5 years ago

if (faceInfo[i].bbox.score < 0.985) continue;

std::cout << "[" << i << "]得分： " << faceInfo[i].bbox.score << std::endl;