Biologically-inspired video enhancement method for robust shape recognition

Saman Poursoltan, Russell Brinkworth, Matthew Sorell

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The way image sequences are encoded by technological systems, that is video, is fundamentally tied to the way in which the human eye and brain interpret images and motion. This includes such aspects as resolution, colour, dynamic range, frame rates and spatial and temporal compression techniques. On the contrary, object identification algorithms are commonly based on single image analysis, such as the extraction of a single video frame from a sequence. This mismatch of, in particular, temporal processing paradigms means that most object analysis algorithms are not well suited to the data with which they are presented. In order to bridge this gap we investigate the temporal preconditioning of video data through a biologically-inspired vision model, based on multi-stage processing analogous to the vision systems of insects. In doing so, we argue that such an approach can lead to improved object identification through the enhancement of object perimeters and the amelioration of lighting and compression artefacts such as shadows and blockiness.

Original languageEnglish
Title of host publication2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013
PublisherIEEE
Number of pages6
ISBN (Electronic)9781467328210, 9781467328197
ISBN (Print)9781467328203, 1467328200
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013 - Sharjah, United Arab Emirates
Duration: 12 Feb 201314 Feb 2013

Publication series

Name2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013

Conference

Conference2013 1st International Conference on Communications, Signal Processing and Their Applications, ICCSPA 2013
CountryUnited Arab Emirates
CitySharjah
Period12/02/1314/02/13

Keywords

  • Human Vision System
  • Shape Recognition
  • video enhancement
  • wavelet transform

Fingerprint Dive into the research topics of 'Biologically-inspired video enhancement method for robust shape recognition'. Together they form a unique fingerprint.

Cite this