Fully Embedding Fast Convolutional Networks on Pixel Processor Arrays

Laurie Bose*, Piotr Dudek, Jianing Chen, Stephen J Carey, Walterio W Mayol-Cuevas

*Corresponding author for this work

Research output: Contribution to conferenceConference Paperpeer-review

Abstract

We present a novel method of CNN inference for pixel processor array (PPA) vision sensors, designed to take advantage of their massive parallelism and analog compute capabilities. PPA sensors consist of an array of processing elements (PEs), with each PE capable of light capture, data storage and computation, allowing various computer vision processing to be executed directly upon the sensor device. The key idea behind our approach is storing network weights "in-pixel" within the PEs of the PPA sensor itself to allow various computations, such as multiple different image convolutions, to be carried out in parallel. Our approach can perform convolutional layers, max pooling, ReLu, and a final fully connected layer entirely upon the PPA sensor, while leaving no untapped computational resources. This is in contrast to previous works that only use a sensor-level processing to sequentially compute image convolutions, and must transfer data to an external digital processor to complete the computation.
We demonstrate our approach on the SCAMP-5 vision system, performing inference of a MNIST digit classification network at over 3000 frames per second and over 93% classification accuracy. This is the first work demonstrating CNN inference conducted entirely upon the processor array of a PPA vision sensor device, requiring no external processing.
Original languageEnglish
Number of pages16
Publication statusPublished - 23 Aug 2020
Event16th European Conference on Computer Vision - Online
Duration: 23 Aug 202028 Aug 2020
https://eccv2020.eu

Conference

Conference16th European Conference on Computer Vision
Abbreviated titleECCV20
Period23/08/2028/08/20
Internet address

Keywords

  • low-level vision
  • PPA
  • CNN
  • vision sensor
  • edge computing

Fingerprint Dive into the research topics of 'Fully Embedding Fast Convolutional Networks on Pixel Processor Arrays'. Together they form a unique fingerprint.

Cite this