A Sparse Convolution Neural Network Accelerator for 3D/4D Point-Cloud Image Recognition on Low Power Mobile Device with Hopping-Index Rule Book for Efficient Coordinate Management

Qiankai Cao, Jie Gu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Scopus citations

Abstract

This work presents the first 3D/4D sparse CNN (SCNN) accelerator for point cloud image recognition on low power devices. A special hopping-index rule book method and efficient data search technique were developed to mitigate the overhead of coordinate management for SCNN. A 65nm test chip for 3D/4D images was demonstrated with 7.09-13.6 TOPS/W power efficiency and state-of-the-art frame rate.

Original languageEnglish (US)
Title of host publication2022 IEEE Symposium on VLSI Technology and Circuits, VLSI Technology and Circuits 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages106-107
Number of pages2
ISBN (Electronic)9781665497725
DOIs
StatePublished - 2022
Event2022 IEEE Symposium on VLSI Technology and Circuits, VLSI Technology and Circuits 2022 - Honolulu, United States
Duration: Jun 12 2022Jun 17 2022

Publication series

NameDigest of Technical Papers - Symposium on VLSI Technology
Volume2022-June
ISSN (Print)0743-1562

Conference

Conference2022 IEEE Symposium on VLSI Technology and Circuits, VLSI Technology and Circuits 2022
Country/TerritoryUnited States
CityHonolulu
Period6/12/226/17/22

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'A Sparse Convolution Neural Network Accelerator for 3D/4D Point-Cloud Image Recognition on Low Power Mobile Device with Hopping-Index Rule Book for Efficient Coordinate Management'. Together they form a unique fingerprint.

Cite this