Reproduce “Spiking Neural Network Enhanced Hand Gesture Recognition Using Low-Cost Single-photon Avalanche Diode Array” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhenya Zang, Xingda Li, David Day Uei Li
A lightweight and cost-effective system for static hand gesture recognition was developed by integrating a low-cost Single-Photon Avalanche Diode (SPAD) array with Spiking Neural Networks (SNNs), achieving 90.8% accuracy and a 20.5% reduction in operations compared to a convolutional neural network while maintaining robustness under ambient light.
Key insights
- Low-cost SPAD arrays can be effectively combined with Spiking Neural Networks (SNNs) for accurate static hand gesture recognition, overcoming the cost barrier of high-end event cameras.
- Training SNNs from scratch with raw photon count data yields models with significantly reduced operational counts (FLOPs) compared to conventional CNNs, showcasing their potential for energy-efficient neuromorphic hardware.
- SNNs, even with simple neuron models like Integrate-and-Fire, exhibit comparable accuracy to CNNs under varying ambient light conditions, indicating robustness for real-world deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.