Low Latency of object detection for spikng neural network

TL;DR

This paper presents a method to enhance the accuracy and reduce the latency of spiking neural networks for object detection by improving conversion techniques and structural modifications, enabling better performance on challenging datasets.

Contribution

The paper introduces a novel conversion and structural adjustment approach that significantly improves SNNs' accuracy and latency for object detection tasks.

Findings

Achieves higher accuracy than previous SNN methods on MS COCO and PASCAL VOC

Reduces latency in SNN-based object detection

Demonstrates advantages of spike signal processing in SNNs

Abstract

Spiking Neural Networks, as a third-generation neural network, are well-suited for edge AI applications due to their binary spike nature. However, when it comes to complex tasks like object detection, SNNs often require a substantial number of time steps to achieve high performance. This limitation significantly hampers the widespread adoption of SNNs in latency-sensitive edge devices. In this paper, our focus is on generating highly accurate and low-latency SNNs specifically for object detection. Firstly, we systematically derive the conversion between SNNs and ANNs and analyze how to improve the consistency between them: improving the spike firing rate and reducing the quantization error. Then we propose a structural replacement, quantization of ANN activation and residual fix to allevicate the disparity. We evaluate our method on challenging dataset MS COCO, PASCAL VOC and our spike dataset. The experimental results show that the proposed method achieves higher accuracy and lower latency compared to previous work Spiking-YOLO. The advantages of SNNs processing of spike signals are also demonstrated.