mtCNN_sysu / README.md
Enderfga's picture
Update README.md
e6ecfeb
metadata
title: MtCNN Sysu
emoji: πŸ“ˆ
colorFrom: gray
colorTo: pink
sdk: gradio
sdk_version: 3.12.0
app_file: app.py
pinned: false
license: openrail

Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks

This repo contains the code, data and trained models for the paper Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks.

Overview

MTCNN is a popular algorithm for face detection that uses multiple neural networks to detect faces in images. It is capable of detecting faces under various lighting and pose conditions and can detect multiple faces in an image.

We have implemented MTCNN using the pytorch framework. Pytorch is a popular deep learning framework that provides tools for building and training neural networks.

Description of file

β”œβ”€β”€ README.md                      # explanatory document
β”œβ”€β”€ get_data.py                    # Generate corresponding training data depending on the input β€œ--net”
β”œβ”€β”€ img                            # mid.png is used for testing visualization effects,other images are the corresponding results.
β”‚   β”œβ”€β”€ mid.png
β”‚   β”œβ”€β”€ onet.png
β”‚   β”œβ”€β”€ pnet.png
β”‚   β”œβ”€β”€ rnet.png
β”‚   β”œβ”€β”€ result.png
β”‚   └── result.jpg
β”œβ”€β”€ model_store                    # Our pre-trained model
β”‚   β”œβ”€β”€ onet_epoch_20.pt
β”‚   β”œβ”€β”€ pnet_epoch_20.pt
β”‚   └── rnet_epoch_20.pt
β”œβ”€β”€ requirements.txt               # Environmental version requirements
β”œβ”€β”€ test.py                        # Specify different "--net" to get the corresponding visualization results
β”œβ”€β”€ test.sh                        # Used to test mid.png, which will test the output visualization of three networks
β”œβ”€β”€ train.out                      # Our complete training log for this experiment
β”œβ”€β”€ train.py                       # Specify different "--net" for the training of the corresponding network
β”œβ”€β”€ train.sh                       # Generate data from start to finish and train
└── utils                          # Some common tool functions and modules
    β”œβ”€β”€ config.py
    β”œβ”€β”€ dataloader.py
    β”œβ”€β”€ detect.py
    β”œβ”€β”€ models.py
    β”œβ”€β”€ tool.py
    └── vision.py

Requirements

  • numpy==1.21.4
  • matplotlib==3.5.0
  • opencv-python==4.4.0.42
  • torch==1.13.0+cu116

How to Install

  • conda create -n env python=3.8 -y
    conda activate env
    
  • pip install -r requirements.txt
    

Preprocessing

  • download WIDER_FACE face detection data then store it into ./data_set/face_detection
  • download CNN_FacePoint face detection and landmark data then store it into ./data_set/face_landmark

Preprocessed Data

# Before training Pnet
python get_data.py --net=pnet
# Before training Rnet, please use your trained model path
python get_data.py --net=rnet --pnet_path=./model_store/pnet_epoch_20.pt
# Before training Onet, please use your trained model path
python get_data.py --net=onet --pnet_path=./model_store/pnet_epoch_20.pt --rnet_path=./model_store/rnet_epoch_20.pt

How to Run

Train

python train.py --net=pnet/rnet/onet #Specify the corresponding network to start training
bash train.sh                        #Alternatively, use the sh file to train in order

The checkpoints will be saved in a subfolder of ./model_store/*.

Finetuning from an existing checkpoint

python train.py --net=pnet/rnet/onet --load=[model path]

model path should be a subdirectory in the ./model_store/ directory, e.g. --load=./model_store/pnet_epoch_20.pt

Evaluate

Use the sh file to test in order

bash test.sh

To detect a single image

python test.py --net=pnet/rnet/onet  --path=test.jpg

To detect a video stream from a camera

python test.py --input_mode=0

The result of "--net=pnet"

The result of "--net=rnet"

The result of "--net=onet"