Open joyhuang9473 opened 7 years ago
This project provide you a method to update multi-task-loss for multi-input source. https://github.com/CongWeilin/mtcnn-caffe
https://github.com/dlunion/mtcnn
基于caffe的mtcnn训练实现,可以训练一个自己的有效的目标检测算法,非常容易非常简单,并且有配套的纯c++版本的mtcnn-light
https://github.com/foreverYoungGitHub/MTCNN
implemented with Caffe, C++ interface.
https://github.com/Tencent/ncnn/issues/52#issuecomment-318555723
手工改了下MTCNN caffemodel转row-major,
https://github.com/ElegantGod/ncnn/blob/master/tools/caffe2ncnn_mtcnn.cpp
The Relationship Between Precision-Recall and ROC Curves http://pages.cs.wisc.edu/~jdavis/davisgoadrichcamera2.pdf
half model shrink the orignal onet by half quantities of parameters
The training data generate process: https://github.com/Seanlinx/mtcnn
Implement MTCNN for face detection and landmark localization in ncnn framework https://github.com/rocking5566/face-detection-MTCNN-ncnn
Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks