To recognize or not to recognize – A database of encrypted images with subjective recognition ground truth

The assessment of very low quality visual data is known to be difficult. In particular, the ability humans recognize encrypted currently impossible determine computationally. human vision research community has widely studied some particular topics, such as image or determination a visibility threshold, while others are still barely researched, specifically content recognition. To this day, there does not exist reliable recognition index that can employed for tasks. order enable study recognition, and in an attempt propose corresponding recognizability index, we build dataset selectively images together with subjective ground-truth about their intelligibility. methods acquisition, setup, protocol, outlier detection, described suggest how calculate score well threshold. performance traditional indices predict assessed on these found inapt estimate content. Contrasting, structure based proposed task shown represent promising starting point further research. facilitate creation foster into its relation system will make database publicly available.