Visual orientation inhomogeneity based scale-invariant feature transform

摘要: Provide the evidence of existence least important visual orientation.Novel algorithm with high efficiency is proposed to detect and describe local feature.Better performance for detection matching, comparable recognition. Scale-invariant feature transform (SIFT) an features in images. In last fifteen years, SIFT plays a very role multimedia content analysis, such as image classification retrieval, because its attractive character on invariance. This paper intends explore new path research by making use findings from neuroscience. We propose more efficient compact scale-invariant detector descriptor simulating orientation inhomogeneity human system. validate that (V-SIFT) can achieve better or at less computation resource time cost various computer vision tasks under real world conditions, matching object work also illuminates wider range opportunities integrating other position-dependent detectors descriptors.