细化语义和强化感知的小目标检测
Small object detection by refining semantics and enhancing percep-tion
袁姮 1王嘉丽 1孟庆姣 1韩荣腾2
作者信息
- 1. 辽宁工程技术大学软件学院,辽宁葫芦岛 125105
- 2. 辽宁工程技术大学工商管理学院,辽宁葫芦岛 125105
- 折叠
摘要
针对在小目标检测过程中因浅层特征语义信息不丰富,导致漏检问题,提出一种多层特征融合改进SSD(single shot multi-box detector)方法.首先在浅层网络中加入深度可分离卷积(depthwise separable convolution,DSC),使用逐通道卷积和逐点卷积强化浅层语义信息;然后将深层网络和浅层网络通过反卷积和空洞卷积的方式细化特征;最后在深层网络中加入注意力机制,增强深层网络对小目标的检测能力.在VOC2007和VOC2012数据集上进行验证,平均检测精度相较于基准算法提高了 5.56％,相较于其他先进算法提升了 4.25％.实验结果表明,提出的细化语义和强化感知方法可以达到提高小目标检测精度的目的.
Abstract
Aiming at the problem of missing detection in the process of small target detection due to insufficient semantic information of shallow features,a multi-layer feature fusion improved single shot multi-box detector(SSD)method is proposed.Firstly,deepwise separable convolution(DSC)is added to the shallow network,and the shallow semantic information is strengthened by channel-by-channel convolution and point-by-point convolution.Then the features of deep network and shallow network are refined utilizing deconvolution and dilation convolution.Finally,the attention mechanism is added to the deep network to enhance the detection ability of small targets.Verified on VOC2007 and VOC2012 data sets,the average detection accuracy is improved by 5.56％compared with the benchmark algorithm and 4.25％compared with other advanced algorithms.The experimental results show that the proposed refined semantics and enhanced perception methods can achieve the purpose of improving the detection accuracy of small targets.
关键词
目标检测/注意力机制/深度可分离卷积(DSC)/SSD算法Key words
object detection/attention mechanism/depthwise separable convolution(DSC)/single shot multi-box detector(SSD)algorithm引用本文复制引用
基金项目
国防预研基金项目(172068)
辽宁省自然科学基金项目(20170540426)
辽宁省教育厅重点基金(LJYL049)
出版年
2024
NETL
NSTL
万方数据