Schlüsselwörter
(Englisch)
|
High Dynamic Range; HDR image; HDR video; HDR 3D; image compression; video compression; HDR subjective quality protocols; HDR objective quality metrics
|
Forschungsprogramme
(Englisch)
|
COST-Action IC1005 - HDRi: The digital capture, storage, transmission and display of real-world lighting
|
Kurzbeschreibung
(Englisch)
|
High Dynamic Range (HDR) imaging is an increasingly popular topic and has been the focus of attention in scientific, technical and artistic communities for several years. It is believed to be the next frontier in imaging similar to transition from gray level to color or 2D to 3D. However, despite many recent developments there is still no standard approach for compression of HDR images and video, nor a way, in which quality is assessed in such imaging modality. Most efforts have been either concentrated on algorithms for conversion of HDR to Low Dynamic Range (LDR), known as tone mapping, or the design of proprietary storage formats for compression of HDR images and video. There is a need for definition of efficient quality metrics and compression formats, that is widely accepted and standardized by international standardization committees such as ISO and ITU. This proposal is to address these shortcomings. On one hand, it aims at setting up a complete quality evaluation pipeline (both subjective protocols and objective metrics) for HDR image and video content, relying on context and environmental parameters, through design of a model for perception of HDR content (image and video) by human subjects. On the other hand, the latter is used to design, implement, and assess the performance of more efficient HDR image and video compression algorithms backward compatible with JPEG image compression and HEVC video compression, respectively, as well as the design of new HDR compression algorithms. Finally, both evaluation methodologies and compression algorithms are extended to 3D HDR content. A public dataset of HDR image and video content together with subjective evaluation scores will be made available to public for advancing the research in this field further.
|
Partner und Internationale Organisationen
(Englisch)
|
AT; BE; BA; CY; CZ; DK; FR; DE; EL; ID; IL; IT; NL; NO; PL; PT; RS; SK; ES; SE; CH; MK; UK
|
Abstract
(Englisch)
|
High Dynamic Range (HDR) imaging is an increasingly popular topic and has been the focus of attention in scientific, technical and artistic communities for several years. It is believed to be the next frontier in imaging similar to transition from gray level to color or 2D to 3D. However, despite many recent developments there is still no standard approach for compression of HDR images and video, nor a way, in which quality is assessed in such imaging modality. Most efforts have been either concentrated on algorithms for conversion of HDR to Low Dynamic Range (LDR), known as tone mapping, or the design of proprietary storage formats for compression of HDR images and video. There is a need for definition of efficient quality metrics and compression formats, which are widely accepted and standardized by international standardization committees such as ISO and ITU. This project addressed these shortcomings. On one hand, it aimed at setting up a complete quality evaluation pipeline (both subjective protocols and objective metrics) for HDR image and video content, relying on context and environmental parameters, through design of a model for perception of HDR content (image and video) by human subjects. On the other hand, the latter was used to design, implement, and assess the performance of more efficient HDR image and video compression algorithms backward compatible with JPEG image compression and HEVC video compression, respectively, as well as the design of new HDR compression algorithms. Finally, both evaluation methodologies and compression algorithms were extended to 3D HDR content. A public dataset of HDR image and video content together with subjective evaluation scores has been made available to public for advancing the research in this field further.
|
Datenbankreferenzen
(Englisch)
|
Swiss Database: COST-DB of the State Secretariat for Education and Research Hallwylstrasse 4 CH-3003 Berne, Switzerland Tel. +41 31 322 74 82 Swiss Project-Number: C12.0081
|
