This work is used for reproduce MTCNN, a Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks.
- You need CUDA-compatible GPUs to train the model.
- Tensorflow 1.2.1
- Python 2.7
- Ubuntu 16.04
- Cuda 8.0
- Download the WIDER Face Dataset. Unzip it and put the
WIDER_train
folder in theprepare_data
folder. - Download the Landmark Dataset. Unzip it and put both the
lfw_5590
andnet_7876
into theprepare_data
folder. - Run
gen_12net_data.py
to generate training data(Face Detection Part) for PNet. - Run
gen_landmark_data.py lfwnet PNet
to generate training data(Face Landmark Detection Part) for PNet. - Run
gen_imglist.py PNet
to merge two parts of training data. - Run
gen_tfrecords.py PNet
to generate tfrecord for PNet. - In
train_models/
, runtrain_net.py PNet
to train PNet. - After training PNet, run
gen_hard_example.py --test_mode PNet
to generate training data(Face Detection Part) for RNet. - Run
gen_landmark_data.py lfwnet RNet
to generate training data(Face Landmark Detection Part) for RNet. - Run
gen_imglist.py PNet
to merge two parts of training data. - Run
gen_tfrecords.py RNet
to generate tfrecords for RNet. - In
train_models/
, runtrain_net.py RNet
to train RNet. - After training RNet, run
gen_hard_example.py --test_mode RNet
to generate training data(Face Detection Part) for ONet. - Run
gen_landmark_data.py lfwnet ONet
to generate training data(Face Landmark Detection Part) for ONet. - Run
gen_imglist.py ONet
to merge two parts of training data. - Run
gen_tfrecords.py ONet
to generate tfrecords for ONet. - In
train_models/
, runtrain_net.py ONet
to train ONet.
-
When training PNet,I merge four parts of data(pos,part,landmark,neg) into one tfrecord,since their total number radio is almost 1:1:1:3.But when training RNet and ONet,I generate four tfrecords,since their total number is not balanced.During training,I read 64 samples from pos,part and landmark tfrecord and read 192 samples from neg tfrecord to construct mini-batch.
-
It's important for PNet and RNet to keep high recall radio.When using well-trained PNet to generate training data for RNet,I can get 14w+ pos samples.When using well-trained RNet to generate training data for ONet,I can get 19w+ pos samples.
-
Since MTCNN is a Multi-task Network,we should pay attention to the format of training data.The format is:
[path to image][cls_label][bbox_label][landmark_label]
For pos sample,cls_label=1,bbox_label(calculate),landmark_label=[0,0,0,0,0,0,0,0,0,0].
For part sample,cls_label=-1,bbox_label(calculate),landmark_label=[0,0,0,0,0,0,0,0,0,0].
For landmark sample,cls_label=-2,bbox_label=[0,0,0,0],landmark_label(calculate).
For neg sample,cls_label=0,bbox_label=[0,0,0,0],landmark_label=[0,0,0,0,0,0,0,0,0,0].
-
Since the training data for landmark is less.I use transform,random rotate and random flip to conduct data augment(the result of landmark detection is not that good).
MIT LICENSE
- Kaipeng Zhang, Zhanpeng Zhang, Zhifeng Li, Yu Qiao , " Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks," IEEE Signal Processing Letter
- MTCNN-MXNET
- MTCNN-CAFFE
- deep-landmark