Surveillance Image Coding for Interpretability Using Importance Maps
BibTeX
@MISC{_surveillanceimage,
author = {},
title = {Surveillance Image Coding for Interpretability Using Importance Maps},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
An image coding technique called importance coding is developed to improve the interpretability versus bit-rate performance for an image compression system. The term interpretability is a subjective image quality measure of the content recognition performance by human observers. Importance coding aims to improve the image interpretability by optimising the amount of useful visual information encoded for a given bit rate. Traditional PSNR scalable compression schemes such as the EZW and SPIHT rely on descending wavelet coefficient magnitudes to prioritise the encoded bit-stream for rate-distortion optimisation. However, for large surveillance imagery, prioritisation of the wavelet coefficients based on the order of importance of contents in an image would be more desirable. An importance map, which provides a systematic approach for the assignment of relative importance to coefficients, is used to aid the prioritisation of the encoded bit-stream according to coefficient importance for interpretability. The importance coding framework discussed in this paper can be incorporated into the JPEG2000 international standard for image coding. Subjective evaluations indicate that importance coding is better than traditional PSNR scalable coders for very low bit-rate content recognition.
Keyphrases
interpretability using importance map surveillance image coding encoded bit-stream importance coding human observer traditional psnr scalable coder subjective evaluation rate-distortion optimisation content recognition performance jpeg2000 international standard low bit-rate content recognition traditional psnr scalable compression scheme wavelet coefficient magnitude coefficient importance large surveillance imagery bit rate relative importance importance map image interpretability subjective image quality measure systematic approach image coding technique term interpretability bit-rate performance image coding useful visual information wavelet coefficient image compression system
