How Good is Too Good? A Subjective Study on Over Enhancement of Images

For a long time different studies have focused on introducing new image enhancement techniques. While these techniques show good performance and are able to increase the quality of images, little attention has been paid how when overenhancement occurs in image. This could possibly be linked fact that current metrics not accurately evaluate enhanced images. In this study we introduce Subjective Enhanced Image Dataset (SEID) which 15 observers asked enhance 30 reference images shown them once at low another high contrast. Observers were instructed point any more will result drop quality. Results there is an agreement between over-enhancement closely similar no matter if contrast or enhanced.