Observer Behavior in Evaluating Images with Region-Specific Distortions: Subjective Scores, Response Time, and Fixation Distribution.

In this study, we aim to investigate whether observers follow a specific pattern when evaluating images that have been distorted at various levels in different areas of the image. In our experiments, three aspects of the rating experience are considered: the subjective score, the response time, and the fixation distribution. Our findings suggest that with the increase in the quality of an image, observers tend to spend more time evaluating their quality. When it comes to images distorted by more than one level of distortion, more than 50% of observers spend more time evaluating the less distorted region of the image, while others do not show any particular bias when performing image quality assessment.