Research on Discrete Fourier Transform Representation for Image Quality Assessment

2014 ◽  
Vol 905 ◽  
pp. 552-555
Author(s):  
Shi Lin Zhao ◽  
Xiu Hua Ji
Keyword(s):  
Fourier Transform ◽  
Quality Assessment ◽  
Discrete Fourier Transform ◽  
Image Quality Assessment ◽  
Structural Similarity ◽  
Simple Calculation ◽  
Assessment Method ◽  
Objective Quality ◽  
Quality Assessment Method ◽  
Blurred Images

Recently the Structural Similarity (SSIM) is proposed, and attracts a lot of attentions for its good performance and simple calculation. By deeply studying the SSIM, we find it fails to measure the badly blurred images. Based on this, we develop an improved objective quality assessment method which is based on Discrete Fourier Transform representation (called as MDFT). Experiment results show the proposed method is more consistent with HVS than SSIM especially for blurred images and fading images.

A New Image Quality Assessment Method Based on SSIM and TV Model

2011 ◽  
Vol 65 ◽  
pp. 542-550
Author(s):  
Lu Lu Pang ◽  
Cong Li Li ◽  
De Ning Qi ◽  
Tao Zou
Keyword(s):  
Image Quality ◽  
Quality Assessment ◽  
Explicit Knowledge ◽  
Image Quality Assessment ◽  
Structural Similarity ◽  
Assessment Method ◽  
Reference Image ◽  
Distorted Image ◽  
Quality Assessment Method ◽  
Reference Images

In this paper, a new image quality assessment method has been proposed in which can judge the quality of images without explicit knowledge of the reference images ,it is based on the SSIM(Structural Similarity) and TV(total variation) model. Firstly, add noises to distorted image to quantitatively determine, it can get the degraded image; secondly, use the improved self-adaptive gradient weights of the TV algorithms to denoising the distorted image, it can get the “fake” reference image, then use the classical SSIM methods to make reference evaluation between the distorted image and the “fake” reference image, after modified, the results is the no reference evaluating indicator. The experiment separated use the standard testing images and the degraded images from the LIVE database to make evaluate experiment, the result show that it is consistent to the result of MOS. This method is no need of reference images, it can use widely.

Sign in / Sign up

Export Citation Format

Share Document