首页|Findings from University of Electronic Science and Technology of China Reveals N ew Findings on Artificial Intelligence (Synthetic Aperture Radar Automatic Targe t Recognition Based On a Simple Attention Mechanism)
Findings from University of Electronic Science and Technology of China Reveals N ew Findings on Artificial Intelligence (Synthetic Aperture Radar Automatic Targe t Recognition Based On a Simple Attention Mechanism)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Research findings on Machine Learning-Artificial Intelligence are discussed in a new report. According to news repor ting originating from Chengdu, People's Republic of China, by NewsRx corresponde nts, research stated, "A simple but effective channel attention module is propos ed for Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR). The ch annel attention technique has shown recent success in improving Deep Convolution al Neural Networks (CNN)." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). Our news editors obtained a quote from the research from the University of Elect ronic Science and Technology of China, "The resolution of SAR images does not su rpass optical images thus information flow of SAR images becomes relatively poor when the network depth is raised blindly leading to a serious gradients explosi on/vanishing. To resolve the issue of SAR image recognition efficiency and ambig uity trade-off, we proposed a simple Channel Attention module into the ResNet Ar chitecture as our network backbone, which utilizes few parameters yet results in a performance gain. Our simple attention module, which follows the implementati on of Efficient Channel Attention, shows that avoiding dimensionality reduction is essential for learning as well as an appropriate cross-channel interaction ca n preserve performance and decrease model complexity. We also explored the One P olicy Learning Rate on the ResNet-50 architecture and compared it with the propo sed attention based ResNet-50 architecture. A thorough analysis of the MSTAR Dat aset demonstrates the efficacy of the suggested strategy over the most recent fi ndings."
ChengduPeople's Republic of ChinaAsi aArtificial IntelligenceAutomatic Target RecognitionEmerging TechnologiesMachine LearningUniversity of Electronic Science and Technology of China
NETL
NSTL
