Research on Method for Constructing High-Dimensional SURF

SURF(Speeded Up Robust Features) is an image descriptor, which is an important method for local features analysis and recognition. SURF is robust, it can do object recognition in images. Meanwhile, SURF uses box filters and integral images for image convolutions to improve the real-time process. This paper presents a constructor method of high-dimensional SURF. Experiment shows that the high-dimensional SURF is more efficiently and accurately in local features description and object recognition than SURF.