Authors: Mingxaun Liu, Jiankai Tang, Haoxiang Li, Jiahao Qi, Siwei Li, Kegang Wang, Yuntao Wang, Hong Chen

Published on: February 07, 2024

Impact Score: 8.15

Arxiv code: Arxiv:2402.04798

Summary

• What is new: First use of SNNs for rPPG, proposing a hybrid model, the Spiking-PhysFormer, designed to cut down power usage.
• Why this is important: Existing ANN-based methods for rPPG require significant computing resources, hindering deployment on mobile devices.
• What the research proposes: A hybrid neural network combining ANNs and SNNs, specifically designed to reduce power consumption without sacrificing performance.
• Results: Achieves a 12.4% reduction in power consumption compared to PhysFormer, with a transformer block that uses 12.2 times less power while maintaining comparable performance.

Technical Details

Technological frameworks used: nan

Models used: Spiking Neural Networks, Artificial Neural Networks, Spiking-PhysFormer

Data used: PURE, UBFC-rPPG, UBFC-Phys, MMPD datasets

Potential Impact

Mobile health monitoring, wearable technology companies, smartphone manufacturers

Want to implement this idea in a business?

We have generated a startup concept here: HeartLink.