Vision

Research output: Chapter in Book/Conference proceedingChapterpeer-review

Abstract

Vision is a sensory modality of primary importance to many animal species. The efficient implementation of visual perception is also one of the main challenges in the design of intelligent robotic systems. This chapter reviews the principles of operation and key features of the early stages of biological vision systems. Following the observation that visual information processing starts in the eye, it reviews several approaches to constructing biomimetic artificial vision systems. It presents devices inspired by the morphology of the insects’ compound eyes, and devices tightly integrating image sensing and processing circuitry. These include silicon integrated circuits mimicking the operation of vertebrate retinas, and bio-inspired systems oriented towards machine vision applications, such as dynamic vision sensors and vision chips with pixel-parallel cellular processor arrays. It elucidates the advantages of the near-sensor processing of the visual information, and potential for future developments of neuromorphic vision sensors.
Original languageEnglish
Title of host publicationLiving machines
Subtitle of host publicationA handbook of research in biomimetics and biohybrid systems
EditorsTony J. Prescott, Nathan Lepora, Paul F.M.J. Verschure
Place of PublicationOxford
PublisherOxford University Press
Chapter14
Pages136-146
Number of pages11
ISBN (Electronic)9780191842702
ISBN (Print)9780199674923
DOIs
Publication statusPublished - 26 Apr 2018

Keywords

  • compound eye
  • reina
  • early vision
  • CMOS image sensor
  • vision chip
  • dynamic vision sensor
  • sensor-level processing
  • pixel-parallel processor array
  • neuromorphic engineering

Fingerprint

Dive into the research topics of 'Vision'. Together they form a unique fingerprint.

Cite this