講演抄録/キーワード |
講演名 |
2008-07-10 11:40
An Insect Vision-based Single-electron Circuit Performing Motion Detection ○Andrew Kilinga Kikombo・Tetsuya Asai・Yoshihito Amemiya(Hokkaido univ.) ED2008-70 SDM2008-89 エレソ技報アーカイブへのリンク:ED2008-70 SDM2008-89 |
抄録 |
(和) |
Nano-electronic devices are promising candidates for the next-generation of low-power LSIs and applications in parallel signal processing LSI systems. So far several fabrication methodologies for single-electron devices have been proposed. However, to apply these devices (which perform on new concepts in relation to the present conventional LSI devices) in functional circuits, we need to come up with new circuit architectures that (i) can accommodate their physical properties and high non-uniformity amongst individual devices (ii) fully utilize their high device density and (iii) provide a solution to their low-fault tolerance operation. One promising solution is to learn from living organisms, particularly insects, to create robust and highly effective signal processors---neuromorphic architectures. In this research, we propose an elementary image processor performing motion detection in incident images. The proposed circuit architecture is based on a widely studied neural correlation model. |
(英) |
Nano-electronic devices are promising candidates for the next-generation of low-power LSIs and applications in parallel signal processing LSI systems. So far several fabrication methodologies for single-electron devices have been proposed. However, to apply these devices (which perform on new concepts in relation to the present conventional LSI devices) in functional circuits, we need to come up with new circuit architectures that (i) can accommodate their physical properties and high non-uniformity amongst individual devices (ii) fully utilize their high device density and (iii) provide a solution to their low-fault tolerance operation. One promising solution is to learn from living organisms, particularly insects, to create robust and highly effective signal processors---neuromorphic architectures. In this research, we propose an elementary image processor performing motion detection in incident images. The proposed circuit architecture is based on a widely studied neural correlation model. |
キーワード |
(和) |
single-electon / neuromorphic circuits / bio-inspired circuits / Beyond CMOS architectures / / / / |
(英) |
single-electon / neuromorphic circuits / bio-inspired circuits / Beyond CMOS architectures / / / / |
文献情報 |
信学技報, vol. 108, no. 122, SDM2008-89, pp. 159-164, 2008年7月. |
資料番号 |
SDM2008-89 |
発行日 |
2008-07-02 (ED, SDM) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
ED2008-70 SDM2008-89 エレソ技報アーカイブへのリンク:ED2008-70 SDM2008-89 |