An Energy-Efficient Matching Accelerator Using Matching Prediction for Mobile Object Recognition

被引：0

作者：

Choi, Seongrim ^{[1
]}

Lee, Hwanyong ^{[1
]}

Nam, Byeong-Gyu ^{[1
]}

机构：

[1] Chungnam Natl Univ, Dept Comp Sci & Engn, 99 Daehak Ro, Daejeon 305764, South Korea

来源：

JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE | 2016年 / 16卷 / 02期

关键词：

Object recognition; object matching; vocabulary tree; matching prediction; matching accelerator;

D O I：

10.5573/JSTS.2016.16.2.251

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

An energy-efficient object matching accelerator is proposed for mobile object recognition based on matching prediction scheme. Conventionally, vocabulary tree has been used to save the external memory bandwidth in object matching process but involved massive internal memory transactions to examine each object in a database. In this paper, a novel object matching accelerator is proposed based on matching predictions to reduce unnecessary internal memory transactions by mitigating nontarget object examinations, thereby improving the energy-efficiency. Experimental results show a 26% reduction in power-delay product compared to the prior art.

引用

页码：251 / 254

页数：4