Affiliation(s):
National Digital Switching System Engineering & Technological R&D Center, Zhengzhou, Henan 450003, China;
moreAffiliation(s): National Digital Switching System Engineering & Technological R&D Center, Zhengzhou, Henan 450003, China; School of Computer Science, Fudan University, Shanghai 201203, China; College of Cyber Science and Engineering, Southeast University, Nanjing 211189, China.;
less
LiChen, FanZhang, Guangwei Xie, Yanzhao Gao, Xiaofeng Qi, Mingqian Sun. S3Det: A fast object detector for remote sensing images based on analog-spiking neural network conversion[J]. Frontiers of Information Technology & Electronic Engineering,in press.https://doi.org/10.1631/FITEE.2400594
@article{title="S3Det: A fast object detector for remote sensing images based on analog-spiking neural network conversion", author="LiChen, FanZhang, Guangwei Xie, Yanzhao Gao, Xiaofeng Qi, Mingqian Sun", journal="Frontiers of Information Technology & Electronic Engineering", year="in press", publisher="Zhejiang University Press & Springer", doi="https://doi.org/10.1631/FITEE.2400594" }
%0 Journal Article %T S3Det: A fast object detector for remote sensing images based on analog-spiking neural network conversion %A LiChen %A FanZhang %A Guangwei Xie %A Yanzhao Gao %A Xiaofeng Qi %A Mingqian Sun %J Frontiers of Information Technology & Electronic Engineering %P %@ 2095-9184 %D in press %I Zhejiang University Press & Springer doi="https://doi.org/10.1631/FITEE.2400594"
TY - JOUR T1 - S3Det: A fast object detector for remote sensing images based on analog-spiking neural network conversion A1 - LiChen A1 - FanZhang A1 - Guangwei Xie A1 - Yanzhao Gao A1 - Xiaofeng Qi A1 - Mingqian Sun J0 - Frontiers of Information Technology & Electronic Engineering SP - EP - %@ 2095-9184 Y1 - in press PB - Zhejiang University Press & Springer ER - doi="https://doi.org/10.1631/FITEE.2400594"
Abstract: Artificial neural networks (ANNs) have made great strides in the field of remote sensing image object detection. However, low detection efficiency and high power consumption have always been a significant bottleneck in remote sensing. Spiking neural networks (SNNs) process information in the form of sparse spikes, creating the advantage of high energy efficiency for computer vision tasks. However, most work has focused on simple classification tasks, and only a few researchers have applied SNNs to object detection in natural images. In this study, we consider the parsimonious nature of biological brains and propose a fast ANN-to-SNN conversion method for remote sensing image detection. We establish a fast sparse model for pulse sequence perception based on group sparse features and conduct transform-domain sparse resampling of the original image to enable fast perception of image features and encoded pulse sequences. In addition, to meet accuracy requirements in relevant remote sensing scenarios, we analyze the transformation error theoretically and propose channel self-decaying weighted normalization to elimi-nate neuron overactivation. We propose S3Det, a remote sensing image object detection model. Our experiments, based on a large publicly available remote sensing dataset, show that S3Det achieves an accuracy performance similar to that of the ANN. Meanwhile, our transformed network is only 24% sparser than the benchmark and consumes only 1.46 J. On the simulation platform, our algorithm improves the integrated inference time 48 times more than the benchmark and consumes a fraction (1/122) of the original algorithm's power.
Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article
Reference
Open peer comments: Debate/Discuss/Question/Opinion
Open peer comments: Debate/Discuss/Question/Opinion
<1>