删除或更新信息,请邮件至freekaoyan#163.com(#换成@)

面向显著性目标检测的SSD改进模型

本站小编 Free考研考试/2022-01-03

余春艳,
徐小丹,
钟诗俊,
福州大学数学与计算机科学学院 ??福州 ??350108
基金项目:福建省产学合作重大项目(2016H6010),福建省自然科学基金(2015J01420),福建省引导性基金(2016Y0060),福建省卫生教育联合攻关计划项目(WKJ2016-2-26)

详细信息
作者简介:余春艳:女,1976年生,副教授,主要研究方向为智能信息处理、虚拟环境与仿真技术、智能算法等
徐小丹:女,1992年生,硕士生,研究方向为图像处理
钟诗俊:男,1994年生,硕士生,研究方向为图像处理
通讯作者:钟诗俊  n160320046@fzu.edu.cn
中图分类号:TP391

计量

文章访问数:1205
HTML全文浏览量:602
PDF下载量:62
被引次数:0
出版历程

收稿日期:2018-01-26
修回日期:2018-07-17
网络出版日期:2018-07-27
刊出日期:2018-11-01

An Improved SSD Model for Saliency Object Detection

Chunyan YU,
Xiaodan XU,
Shijun ZHONG,
College of Mathematics and Computer Science, Fuzhou University, Fuzhou 350108, China
Funds:The Major Project in Industry-university Cooperation of Fujian Province (2016H6010), The Natural Science Foundation of Fujian Province (2015J01420), The Guiding Found of Fujian Province (2016Y0060), The Health-Education Joint Project of Fujian Province (WKJ2016-2-26)


摘要
摘要:传统显著性目标检测方法常假设只有单个显著性目标,其效果依赖显著性阈值的选取,并不符合实际应用需求。近来利用目标检测方法得到显著性目标检测框成为一种新的解决思路。SSD模型可同时精确检测多个不同尺度的目标对象,但小尺寸目标检测精度不佳。为此,该文引入去卷积模块与注意力残差模块,构建了面向多显著性目标检测的DAR-SSD模型。实验结果表明,DAR-SSD检测精度显著高于SOD模型;相比原始SSD模型,在小尺度和多显著性目标情形下性能提升明显;相比MDF和DCL等深度学习框架下的方法,也体现了复杂背景情形下的良好检测性能。
关键词:目标检测/
显著性目标检测/
去卷积/
注意力残差
Abstract:Traditional saliency object detection methods, assuming that there is only one salient object, is not conductive to practical application. Their effects are dependent on saliency threshold. Object detection model provides a kind of new solutions. SSD can accurately detect multi-objects with different scales simultaneously, except for small objects. To overcome this drawback, this paper presents a new multi- saliency objects detection model, DAR-SSD, appending a deconvolution module embedded with an attention residual module. Experiments show that DAR-SSD achieves a higher detection accuracy than SOD. Also, it improves detection performance for multi- saliency objects on small scales, compared with original SSD, and it has an advantage over complicated background, compared with MDF and DCL, which also are deep model based methods.
Key words:Object detection/
Saliency object detection/
Deconvolutional/
Attention residual



PDF全文下载地址:

https://jeit.ac.cn/article/exportPdf?id=06356e2e-209e-4122-86b8-ec0d2e716708
相关话题/智能 福州大学 计算机 环境 教育