抄録
It has been computationally clarified that how an image is observed affects how the human vision system perceives sharpness. To do this, we used a sharpness evaluation method based on the cooperative human vision model with image observation characteristics similar to those in human vision and examined the effect of the model's mechanism for observing an image on its sharpness metric performance. The evaluated images were continuous-tone generalized knife-edge, their edge-enhanced, and their digital halftone ones. The following results were obtained: (1) the vision model reproduced accommodation characteristics during subjective sharpness evaluation; (2) the sharpness evaluation metric with the image observing mechanism had a superior evaluation performance to the metric without it. These results suggest that the adaptive change in the image observation characteristic for evaluated images prevents the evaluation performance from deteriorating with evaluation conditions and that similar phenomena occur in the sharpness evaluation process in human vision.
