Home

Awesome

PyramidBox: A Context-assisted Single Shot Face Detector.##

A PyTorch Implementation of PyramidBox

Description

I train pyramidbox with pytorch and the result approaches the original paper result,the pretrained model can be downloaded in vgg,the final model can be downloaded in Pyramidbox.the AP in WIDER FACE as following:

Easy MAPMedium MAPhard MAP
origin paper0.9600.9480.888
this repo0.9480.9380.880

the AP in AFW,PASCAL,FDDB as following:

AFWPASCALFDDB
99.6599.020.983

the gap is small with origin paper,I train 120k batch_size 4 which is different from paper,which maybe cause the gap,if you have more gpu ,the final result maybe better.

Requirement

Prepare data

  1. download WIDER face dataset
  2. modify data/config.py
  3. python prepare_wider_data.py

Train

python train.py --batch_size 4  
		--lr 5e-4

Evalution

according to yourself dataset path,modify data/config.py

  1. Evaluate on AFW.
python tools/afw_test.py
  1. Evaluate on FDDB
python tools/fddb_test.py
  1. Evaluate on PASCAL face
python tools/pascal_test.py
  1. test on WIDER FACE
python tools/wider_test.py

Demo

you can test yourself image

python demo.py

Result

<div align="center"> <img src="https://github.com/yxlijun/Pyramidbox.pytorch/blob/master/tmp/gsmarena_001.jpg" height="250px" alt="demo" > <img src="https://github.com/yxlijun/Pyramidbox.pytorch/blob/master/tmp/0_Parade_marchingband_1_488.jpg" height="200px" alt="demo" > </div>

References