Meta-analysis of distortions effects on quality of digital images using standard IQA datasets

Abstract

The research on image quality assessment has been an area of concern for the researchers from last many years. Images, during their life cycle, passes through several phases including acquisition, compression and restoration, for ease of transmission and storage, encryption and decryption for better security. During their processing, the quality of images is affected by many distortions, including compression artifacts, noise, bandwidth loss, etc., that are not present in the original images. The main emphasis of this research work is to analyze the effect of such artifacts on the quality of images. Further, a brief description of available image quality datasets, along with the popular quality evaluation metrics, is also presented. From experimental analysis in the present work, it is observed that when the images are degraded by impurities during their processing phases, then their quality degenerates by a relatively high degree, varying from minimum 1% to maximum of 55% under different sets of distortions, when measured on standard IQA metrics such as MSE, PSNR, PFE, SSIM, etc. This study will motivate the future researchers to have an elaborate quality assessment before employing these images for practical applications like medical sciences, remote sensing, etc.

Subject Classification:

• 68Pxx
• 68Mxx

Keywords:

• Image artifacts
• IQA datasets
• Objective quality metrics
• Subjective quality metrics
• Objective quality metrics