Summary
International Technical Conference on Circuits/Systems, Computers and Communications
2008
Session Number:F5
Session:
Number:F5-4
A Study on VQ based Compression of High Dynamic Range Images
Nagisa Sugiyama, Masahiro Okuda,
pp.-
Publication Date:2008/7/7
Online ISSN:2188-5079
Summary:
An encoding method of the high dynamic range (HDR) images based on Vector Quantization is discussed. The HDR images have much higher dynamic ranges compared with the low dynamic range (LDR) images. Due to its high dynamic range, the size is often huge, thus development for functional compression is needed. On the other hand, we often need to convert them into the LDR images because the existing output devices cannot directly output the dynamic ranges of HDR images. Our goal is to develop a HDRI compression method that minimizes the error of the tone-mapped LDR images. Therefore we applied the adjustment of scale and the evaluation of the error in the log domain. Our method improves a compression performance, compared to methods that directly apply the conventional vector quantization to the HDR images.