Comparison of SSC to human perception

The perception of objects is our eye's everyday task. It is, however a nontrivial one: Attempts to perform this task artificially still struggle with a number of problems, reflected in a low reliability of the methods and comparably high resource demands. For the recognition of objects we have recently designed an efficient and reliable solid-state physics motivated algorithm that outperforms concurrent approaches by far. In the present study, we investigate to what extent our algorithm coincides with the human visual object perception. In order to assess the ways humans perceive objects, we use eye-tracking, and then compare emerging characteristics with those from machine-learning approach.