CASIA OpenIR  > 模式识别国家重点实验室  > 图像与视频分析
Improved Single Shot Object Detector Using Enhanced Features and Predicting Heads
Zhao X(赵旭)1,2; Zhao CY(赵朝阳)1,2; Zhu YS(朱优松)1,2; Tang M(唐明)1,2; Wang JQ(王金桥)1,2
2018-09
Conference Name2018 IEEE Fourth International Conference on Multimedia Big Data (BigMM)
Conference Date2018-09-13~16
Conference Place中国,西安
Abstract

Object detection attracts much attention for its great value in theories and applications. The one-stage single shot object detectors outperform the two-stage methods in running speed with a comparable performance. In this paper, we propose three novel strategies, to further improve the performances of single shot detector without sacrificing their runtime efficiency. Firstly, we design the multi-scale context aggregation module to embeds the context information into the learned features. Secondly, we design the multi-path predicting head, which decouples the network layers and can easily learn the effective receptive fields of different aspect ratios, to detect objects of various aspect ratios better. Thirdly, we adopt a top-down feature map pyramid to detect objects using features of different semantic powers and resolutions. Sufficient ablation experiments are conducted to prove the efficiency of the proposed methods. We design a one-stage single detector named as ISSD, using the three strategies.  Experimental results on PASCAL VOC 2007 and 2012 shows ISSD achieves the new state-of-the-art on accuracy with the comparable running speed. 

Keyword目标检测
MOST Discipline Catalogue工学::计算机科学与技术(可授工学、理学学位)
DOI10.1109/BigMM.2018.8499089
URL查看原文
Indexed ByEI
Language英语
Citation statistics
Document Type会议论文
Identifierhttp://ir.ia.ac.cn/handle/173211/23596
Collection模式识别国家重点实验室_图像与视频分析
Corresponding AuthorTang M(唐明)
Affiliation1.中国科学院自动化研究所
2.中国科学院大学
First Author AffilicationInstitute of Automation, Chinese Academy of Sciences
Corresponding Author AffilicationInstitute of Automation, Chinese Academy of Sciences
Recommended Citation
GB/T 7714
Zhao X,Zhao CY,Zhu YS,et al. Improved Single Shot Object Detector Using Enhanced Features and Predicting Heads[C],2018.
Files in This Item: Download All
File Name/Size DocType Version Access License
AS687362573074433154(1007KB)会议论文 开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Zhao X(赵旭)]'s Articles
[Zhao CY(赵朝阳)]'s Articles
[Zhu YS(朱优松)]'s Articles
Baidu academic
Similar articles in Baidu academic
[Zhao X(赵旭)]'s Articles
[Zhao CY(赵朝阳)]'s Articles
[Zhu YS(朱优松)]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Zhao X(赵旭)]'s Articles
[Zhao CY(赵朝阳)]'s Articles
[Zhu YS(朱优松)]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: AS6873625730744331540891416439_content_1 (1).pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.