Performance of Texture Compression Algorithms in Low-Latency Computer Vision Tasks

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

17 Downloads (Pure)

Abstract

Deep learning has been successfully used for computer vision tasks, but its high computational cost limits the adoption in lightweight devices such as camera sensors. For this reason, many low-latency vision systems offload the inference computation to a local server, requiring fast (de)compression of the source images. Texture compression is a compelling alternative to existing compression schemes, such as JPEG or HEVC, due to its low decoding overhead, straightforward parallelization, robustness, and a fixed compression ratio. In this paper, we study the impact of lightweight bounding box-based texture compression algorithms, BC1 and YCoCg-BC3, on the accuracy of two computer vision tasks: object detection and semantic segmentation. While JPEG achieves superior per-pixel error rate, the YCoCg-BC3 encoding can provide comparable vision accuracy. The BC1 encoding results in significant degradation of vision performance. However, by retraining the FasterSeg teacher network with a BC1-compressed dataset, we reduced its segmentation mIoU loss from 2.7 to 0.5 percent. Thus, both BC1 and YCoCg-BC3 encoders are suitable for use in low latency vision systems, since they both achieve significantly higher encoding speed than JPEG and their decoding overhead is negligible.

Original languageEnglish
Title of host publicationProceedings of the 2021 9th European Workshop on Visual Information Processing, EUVIP 2021
EditorsA. Beghdadi, F. Alaya Cheikh, J.M.R.S. Tavares, A. Mokraoui, G. Valenzise, L. Oudre, M.A. Qureshi
PublisherIEEE
Number of pages6
ISBN (Electronic)9781665432306
ISBN (Print)9781665432313
DOIs
Publication statusPublished - 2021
Publication typeA4 Article in conference proceedings
EventEuropean Workshop on Visual Information Processing - Paris, France
Duration: 23 Jun 202125 Jun 2021

Publication series

NameEuropean Workshop on Visual Information Processing
ISSN (Print)2164-974X
ISSN (Electronic)2471-8963

Conference

ConferenceEuropean Workshop on Visual Information Processing
Country/TerritoryFrance
CityParis
Period23/06/2125/06/21

Keywords

  • Computer Vision
  • Image Compression
  • Low Latency
  • Texture Compression

Publication forum classification

  • Publication forum level 1

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Information Systems
  • Signal Processing

Fingerprint

Dive into the research topics of 'Performance of Texture Compression Algorithms in Low-Latency Computer Vision Tasks'. Together they form a unique fingerprint.

Cite this