A liveness detection method for face recognition based on optical flow field

Abstract

It is a common spoof to use a photograph to fool face recognition algorithm. In light of differences in optical flow fields generated by movements of two-dimensional planes and three-dimensional objects, we proposed a new liveness detection method for face recognition. Under the assumption that the test region is a two-dimensional plane, we can obtain a reference field from the actual optical flow field data. Then the degree of differences between the two fields can be used to distinguish between a three-dimensional face and a two-dimensional photograph. Empirical study shows that the proposed approach is both feasible and effective.