SI-AGAN

Attention-guided gan for synthesizing IR images

This repository contains the Tensorflow code for “Pedestrian Gender Recognition by Style Transfer of Visible-Light Image to Infrared-Light Image Based on an Attention-Guided Generative Adversarial Network“. This code is based on the TensorFlow implementation of AGGAN provide by AlamiMejjati.

I am working on updating to Tensorflow 2.0, and when it is updated, I will add a link to the new repository.

Dataset

  1. ReGDB Database
  2. SYSU-MM01 Database

Training

  1. revise .json file in configs folder

  2. make .csv file using create_cyclegan_dataaset.py

  3. start training

    python main.py --to_train=1 --log_dir=./output/##/## --config_filename=./configs/##.json

Also, you can check the tensorbaord

tensorboard --port=#### --logdir=./output/##/##/#timestamp#

Testing

python main.py --to_train=0 --log_dir=./output/##/## --config_filename=./configs/##.json --checkpoint_dir=./output/##/##/#timestamp#

GitHub

View Github