Subjective MSE Measures

Marmolin, Hans

doi:10.1109/tsmc.1986.4308985

Cited by 119 publications

(53 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some authors expressed their doubt whether their measure was applicable to evaluate de{interlacing. One alternative, the`subjective M S E ', that we experimented with [48] did not lead to dierent conclusions than the common M S E . We therefore see no good alternative yet for the much criticized M S E .…”

Section: Discussionmentioning

confidence: 90%

Deinterlacing-an overview

1998

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 90%

Deinterlacing-an overview

1998

View full text Add to dashboard Cite

“…Some more complex measures that use frequency weighting to take advantage of the CSF have also been used with some success [25], [26], [28], [29], [39]. Of interest is a technique that measures a number of both local and global properties of the error to obtain a quantitative picture quality score (PQS) [27].…”

Section: B Hvs Properties Used In Image Compressionmentioning

confidence: 99%

“…Therefore, the compression noise added by JPEG varies spatially and takes advantage of the CSF and visual masking to some extent [20], [51]. This can also account for the good subjective performance of WPSNR over PSNR [25], [26], [28], [29], as most compression schemes will place the largest errors in the highest frequency areas of an image. These errors are then likely to be highly correlated to the image area, i.e., be high frequency also, and so will receive a low weighting from WPSNR.…”

Section: B a Natural Imagementioning

confidence: 99%

A wavelet visible difference predictor

Bradley

1999

IEEE Trans. on Image Process.

131

View full text Add to dashboard Cite

Abstract-In this paper, we describe a model of the human visual system (HVS) based on the wavelet transform. This model is largely based on a previously proposed model, but has a number of modifications that make it more amenable to potential integration into a wavelet based image compression scheme. These modifications include the use of a separable wavelet transform instead of the cortex transform, the application of a wavelet contrast sensitivity function (CSF), and a simplified definition of subband contrast that allows us to predict noise visibility directly from wavelet coefficients. Initially, we outline the luminance, frequency, and masking sensitivities of the HVS and discuss how these can be incorporated into the wavelet transform. We then outline a number of limitations of the wavelet transform as a model of the HVS, namely the lack of translational invariance and poor orientation sensitivity. In order to investigate the efficacy of this wavelet based model, a wavelet visible difference predictor (WVDP) is described. The WVDP is then used to predict visible differences between an original and compressed (or noisy) image. Results are presented to emphasize the limitations of commonly used measures of image quality and to demonstrate the performance of the WVDP. The paper concludes with suggestions on how the WVDP can be used to determine a visually optimal quantization strategy for wavelet coefficients and produce a quantitative measure of image quality.Index Terms-Human visual system, image compression, image quality, visible difference prediction, wavelet transform.

show abstract

“…Objective measures by computer programs whose evaluations are in close agreement with human judgment have been extensively studied in the past. Early work about objective measures characterized the similarity of two images of same size using peak signal to noise ratio (PSNR) and mean squared errors (MSE) [5]. A structural similarity framework [6], called SSIM, was proposed based on the assumption that human vision system (HVS) is highly adapted for extracting statistic structural information.…”

Section: Introductionmentioning

confidence: 99%