A Discriminative Binary Descriptor Built on Further Mining Marginal Information

Through further analyzing sampling-pattern characteristics of BRISK and FREAK,we find that both sampling-point density and degree of overlapping have an influence on the specificity of descriptor.The two factors could appropriately be tuned to design an optimized sampling pattern,and map it onto the local area of keypoint with right orientation.A coarse descriptor,built by testing sampling points selected randomly on the sampling pattern,is used to learn a fine descriptor from training data.Results based on experiments of performance evaluation under two kinds testing environments have shown that the proposed binary descriptor outperforms the others.The good effectiveness of applying the proposed descriptor into application of 3D construction has proved that the proposed descriptor is robust to variety of image transformations,as well as performs well in real-time applications.