Depth-of-field effect in subjective and objective evaluation of image quality

Abstract

Depth of field, as an important image feature, is always ignored when assessing the quality of photographs. The current study compared the subjective image quality evaluation with three NR image quality algorithms: BIQI, BRISQUE, and NIQE on the performance of scoring photographs with different levels of depth of field. The results showed that a large depth of field led to higher image quality in BRISQUE and depth of field did not influence the performance of the other two objective metrics and human perception. We also concluded that objective metrics such as BRISQUE, NIQE, BIQI underestimated the image quality of photographs compared with human perception.