A Hybrid Neural Coding Approach for Pattern Recognition with Spiking Neural Networks

This repository contains the implementation of the paper A Hybrid Neural Coding Approach for Pattern Recognition with Spiking Neural Networks published at IEEE Transactions on Pattern Analysis Machine Intelligence (TPAMI).

Dataset

CIFAR-10

Dependencies

• python 3.9.18
• Pytorch 1.13.0
• Spikingjelly 0.0.0.12.

Reducibility

• Pre-train the ANN.

  python ./CIFAR10/ANN_baseline/cifar10_vgg16_base_model.py
  python ./CIFAR10/ANN_baseline/cifar10_resNet20_base_model.py
  

• Training Hybrid Coding framework.

  python ./CIFAR10/Hybrid_coding/cifar10_main_vgg16.py
  python ./CIFAR10/Hybrid_coding/cifar10_main_res20.py
  

Citation

@article{chen2024hybrid,
  title={A hybrid neural coding approach for pattern recognition with spiking neural networks},
  author={Chen, Xinyi and Yang, Qu and Wu, Jibin and Li, Haizhou and Tan, Kay Chen},
  journal={IEEE Transactions on Pattern Analysis \& Machine Intelligence},
  volume={46},
  number={05},
  pages={3064--3078},
  year={2024},
  publisher={IEEE Computer Society}
}