A study on the usability of opinion-unaware no-reference natural image quality metrics in the context of medical images

M. Outtas 1 L. Zhang 1 O. Deforges 1 W. Hamidouche 1 A. Serir C. Cavaro-Menard 2
1 IETR - Institut d'Electronique et de Télécommunications de Rennes
2 LARIS - Laboratoire angevins de recherche en ingénierie des systèmes (EA 7315)
Abstract : In this paper, we firstly study the possible usability of two state-of-the-art no-reference image quality assessment (IQA) metrics (NIQE and BIQES) originally designed for natural image, in the context of medical ones. We also proposed a modified NIQE, called NIQE-K, inspired by some BIQES features and more adapted to medical images. This study, based on evaluating concurrently the three methods, first encompasses tests conducted on natural IQA database (LIVE-Release2 and CSIQ). The second experiment is conducted on an ultrasound image with noise distortions. The last experiment includes tests on Magnetic Resonance images with compression distortions analyzed with quality scores evaluated by radiologists. © 2016 IEEE.
Keywords : BIQES Image quality assessment Medical images NIQE NR-IQA
Type de document :
Communication dans un congrès
2016 International Symposium on Signal, Image, Video and Communications, ISIVC 2016, Nov 2016, Tunis, Tunisia. 〈10.1109/ISIVC.2016.7894006〉
Liste complète des métadonnées
https://hal-univ-rennes1.archives-ouvertes.fr/hal-01622526
Contributeur : Laurent Jonchère <>
Soumis le : mardi 24 octobre 2017 - 14:34:32
Dernière modification le : mercredi 16 mai 2018 - 11:23:51

Identifiants

Collections

IETR | UNIV-NANTES | SUP_IETR | CENTRALESUPELEC | UR1-UFR-ISTIC | IETR-VAADER | UNIV-RENNES

Citation

M. Outtas, L. Zhang, O. Deforges, W. Hamidouche, A. Serir, et al.. A study on the usability of opinion-unaware no-reference natural image quality metrics in the context of medical images. 2016 International Symposium on Signal, Image, Video and Communications, ISIVC 2016, Nov 2016, Tunis, Tunisia. 〈10.1109/ISIVC.2016.7894006〉. 〈hal-01622526〉

Exporter

BibTeX TEI DC DCterms EndNote

Partager

Métriques

Consultations de la notice

92