Quality index evaluation of videos based on fuzzy interface system

What is it about?

Objective image quality assessment (IQA) is a challenge facing digital image and video processing systems because image quality is distorted during various applications, including restoration, compression, storage and transmission. Therefore, this study proposes a new methodology based on a fuzzy interface system called quality evaluation system (QES) to measure the total quality index of input videos with many distorted situations. Nine well-known quality metrics (PSNR, VSNR, WSNR, SSIM, MSSIM, UQI, VIF, IFC, and NQM) were used as inputs for three fuzzy logic controller systems, and their outputs were set as inputs to another fuzzy logic controller system to obtain the total quality index (TQI) of the input video. This process contributes to obtain clear performance of the quality index of the input video despite the failure of some IQA methods in providing quality performance of the input video in some situations. The proposed QES is tested on four videos captured with different digital cameras. Each video is divided into three sets based on the distortion types at different noise levels (Gaussian noise, Poisson noise, and blurring). Furthermore, we evaluated the proposed QES on three largest well-known databases (TID2008, TID2013 and LIVE) to improve the experimental results, which indicate that the proposed QES leads to significant improvement and outperforms other IQA methods used in this study. In addition, we used wavelet decomposition and image de-noising to enhance the standard Eulerian video magnification (EVM) technique. The proposed QES was also used to prove that our magnification system has better magnification quality index than other magnification techniques.

Featured Image