Gestalt Interest Points for Image Description in Weight-Invariant Face Recognition

In this work, we propose two improvements of the Gestalt Interest Points (GIP) algorithm for the recognitionof faces of people that have underwent signi cant weight change. The basic assumption is that some interestpoints contribute more to the description of such objects than others. We assume that we can eliminate certaininterest points to make the whole method more e cient while retaining our classi cation results. To nd outwhich gestalt interest points can be eliminated, we did experiments concerning contrast and orientation of facefeatures. Furthermore, we investigated the robustness of GIP against image rotation. The experiments showthat our method is rotational invariant and { in this practically relevant forensic domain { outperforms thestate-of-the-art methods such as SIFT, SURF, ORB and FREAK.