On Edge Image Processing Acceleration with Low Power Neuro-Memristive Segmented Crossbar Array Architecture

Publikation: Beitrag in FachzeitschriftForschungsartikelBeigetragenBegutachtung

Beitragende

  • Nikolaos Vasileiadis - , Democritus University of Thrace, Demokritos National Centre for Scientific Research (Autor:in)
  • Vasileios Ntinas - , Democritus University of Thrace (Autor:in)
  • Panagiotis Karakolis - , Demokritos National Centre for Scientific Research (Autor:in)
  • Panagiotis Dimitrakis - , Demokritos National Centre for Scientific Research (Autor:in)
  • Georgios Ch Sirakoulis - , Democritus University of Thrace (Autor:in)

Abstract

Computational acceleration for image processing tasks on the edge is becoming increasingly important for many applications. This work presents a new neuro-inspired architecture which incorporates in-memory computing properties for image processing complex computational tasks in addition to analog memory. The proposed architecture was based on segmented crossbar topology, which, as it turns out, reduces many phenomena that affect the performance on such systems. The extended architectural capabilities of this structure were also tested in a systematic analysis that was performed on a novel depth map extraction application from a single defocused image. All results were validated through Spice simulations using a novel moving barrier mem-ristor model.

Details

OriginalspracheEnglisch
Seiten (von - bis)173-199
Seitenumfang27
Fachzeitschrift International journal of unconventional computing : non-classical computation and cellular automata
Jahrgang17
Ausgabenummer3
PublikationsstatusVeröffentlicht - 2022
Peer-Review-StatusJa
Extern publiziertJa

Externe IDs

ORCID /0000-0002-2367-5567/work/168720249

Schlagworte

ASJC Scopus Sachgebiete

Schlagwörter

  • convolution engine, depth map, Edge computing, in-memory computing, IoT, memristor, neuromorphic accelerator, segmented crossbar

Bibliotheksschlagworte