Prediction of perceived image quality by mobile device users

Abstract

This paper provides an overview of the PhD research of perceived image quality (PIQ) assessment, which has a great potential for image processing in many applications. The outcome of this research can be used by smartphone vendors in order to shorten “time to market,” or by any image quality experts, particularly in the academia, to reduce the time and cost of PIQ assessment of smartphones with a small high-definition display. The method can be implemented in a software if it is embedded in the social network websites to predict and improve the image quality as perceived by the users in real time. The research was based on no reference (NR) PIQ evaluation experiments combined with the VIQET software application. The experiments include human visual tests (HVTs) for subjective image quality assessment. The results of HVTs analysis identify the most effective image quality attributes for perceived image quality. The VIQET software tool was calibrated according to the HVTs scores.