DC Field | Value | Language |
---|---|---|
dc.contributor.author | Han, Donghyeon | ko |
dc.contributor.author | LEE, Jinsu | ko |
dc.contributor.author | Lee, Jinmook | ko |
dc.contributor.author | Choi, Sungpill | ko |
dc.contributor.author | Yoo, Hoi-Jun | ko |
dc.date.accessioned | 2019-04-15T14:37:01Z | - |
dc.date.available | 2019-04-15T14:37:01Z | - |
dc.date.created | 2018-12-19 | - |
dc.date.created | 2018-12-19 | - |
dc.date.created | 2018-12-19 | - |
dc.date.issued | 2018-05 | - |
dc.identifier.citation | IEEE International Symposium on Circuits & Systems | - |
dc.identifier.uri | http://hdl.handle.net/10203/254224 | - |
dc.description.abstract | A low-power online deep neural network (DNN) training processor is proposed for a real-time object tracking in mobile devices. For a real-time object tracking, a homogeneous core architecture is proposed to achieve 1.33 x higher throughput than previous DNN training processor. To reduce the external memory access (EMA), a binary feedback alignment (BFA) algorithm and an integral run-length compression (iRLC) decoder are proposed. While the BFA reduces the EMA by 11.4% compared to the conventional back-propagation approach, the iRLC decoder achieves 29.7% EMA reduction without throughput degradation. Finally, a dropout controller is proposed and achieves 43.9% power reduction through clock-gating. Implemented with 65 nm CMOS technology, the 4.4 mm(2) DNN training processor achieves 141.1 mW power consumption at 30.4 frames-per-second (fps) real-time object tracking in mobile devices. | - |
dc.language | English | - |
dc.publisher | IEEE International Symposium on Circuits & Systems | - |
dc.title | A 141.4 mW Low-Power Online Deep Neural Network Training Processor for Real-time Object Tracking in Mobile Devices | - |
dc.type | Conference | - |
dc.identifier.wosid | 000451218702083 | - |
dc.identifier.scopusid | 2-s2.0-85057072319 | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | IEEE International Symposium on Circuits & Systems | - |
dc.identifier.conferencecountry | IT | - |
dc.identifier.conferencelocation | Firenze Fiera Congress and Exhibition Center | - |
dc.identifier.doi | 10.1109/ISCAS.2018.8351398 | - |
dc.contributor.localauthor | Yoo, Hoi-Jun | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.