Napredna pretraga

Pregled bibliografske jedinice broj: 872215

Sensitivity of Tone Mapped Image Quality Metrics to Perceptually Hardly Noticeable Differences


Banić, Nikola; Lončarić, Sven
Sensitivity of Tone Mapped Image Quality Metrics to Perceptually Hardly Noticeable Differences // Proceedings of The Fifth Croatian Computer Vision Workshop (CCVW 2016)
Osijek, Hrvatska, 2016. str. 15-18 (poster, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Sensitivity of Tone Mapped Image Quality Metrics to Perceptually Hardly Noticeable Differences

Autori
Banić, Nikola ; Lončarić, Sven

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of The Fifth Croatian Computer Vision Workshop (CCVW 2016) / - , 2016, 15-18

Skup
Fifth Croatian Computer Vision Workshop

Mjesto i datum
Osijek, Hrvatska, 11.10.2016

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
High dynamic range ; objective image quality assessment ; FSITM ; low dynamic range ; TMQI ; TMQI-II ; tone mapping

Sažetak
As high dynamic range images are being used more widely, the need for good tone mapping operators (TMOs) i.e. methods for their conversion to low dynamic range images rises as well. In evaluation of results of TMOs objective image quality metrics are often used for practical reasons. Since these metrics only approximate perceptual evaluation, they are sometimes too sensitive to perceptually unimportant details. In this paper such sensitivities of three recent tone mapped image quality metrics are compared: TMQI, TMQI-II, and FSITM. These metrics have been chosen because they are the most appropriate objective quality metrics for the problem of tone mapping. The comparison is performed by using specifically designed tone mapped images to check the measures' susceptibility to perceptually unnoticeable changes in brightness of the resulting image. It is shown that while values of TMQI and FSITM are only slightly affected by such changes, the recent TMQI-II can obtain significantly different values, which brings into question its ability perform a fair TMO comparison. The results are presented and discussed.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo



POVEZANOST RADA


Projekt / tema
036-0362214-1989 - Inteligentne metode obrade i analize slika (Sven Lončarić, )

Ustanove
Fakultet elektrotehnike i računarstva, Zagreb