Ke Xu, Yu Li, Bin Han, Xiao Zhang, Xin Liu, Jisong Ai
{"title":"A Low-power Computer Vision Engine for Video Surveillance","authors":"Ke Xu, Yu Li, Bin Han, Xiao Zhang, Xin Liu, Jisong Ai","doi":"10.1109/CICTA.2018.8705947","DOIUrl":null,"url":null,"abstract":"this paper presents the design and VLSI implementation of a CVE (Computer Vision Engine) for real-time video analysis. It offloads CPU/GPU for the power-hungry computation for various vision tasks such as face detection, object detection, motion tracking, etc. The design features 22 computation kernels and is divided into three main categories. The proposed CVE is integrated in a smart video surveillance SoC (System on Chip) and fabricated with TSMC 28nm technology. The total hardware costs are 392K gates and 75.5 KB memory. The measured results show that the design is able to achieve $1920\\times1080$ 30fps real-time video analysis when running at 400MHz. The total power consumption is 20mW and 0.32nJ/pixel of energy efficiency.","PeriodicalId":186840,"journal":{"name":"2018 IEEE International Conference on Integrated Circuits, Technologies and Applications (ICTA)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2018 IEEE International Conference on Integrated Circuits, Technologies and Applications (ICTA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CICTA.2018.8705947","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
this paper presents the design and VLSI implementation of a CVE (Computer Vision Engine) for real-time video analysis. It offloads CPU/GPU for the power-hungry computation for various vision tasks such as face detection, object detection, motion tracking, etc. The design features 22 computation kernels and is divided into three main categories. The proposed CVE is integrated in a smart video surveillance SoC (System on Chip) and fabricated with TSMC 28nm technology. The total hardware costs are 392K gates and 75.5 KB memory. The measured results show that the design is able to achieve $1920\times1080$ 30fps real-time video analysis when running at 400MHz. The total power consumption is 20mW and 0.32nJ/pixel of energy efficiency.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
