A novel two stage image recognition method

Abstract

In this paper, a novel algorithm for image recognition, consisting of two stages: coarse recognition and fine recognition, is presented. For coarse recognition, a new gray-spatial histogram is proposed, which incorporates spatial information with gray-scale compositions without sacrificing the robustness of traditional gray histograms. For fine recognition, a new wavelet set, called directional wavelets, is obtained from the exponential wavelet family. And then an approach to directional wavelet transform-based recognition using feature weighting is proposed. The weighted directional wavelet coefficients represent the directionality and multi-scale characteristics of images. The combination of coarse and fine recognition steps reduces the computational cost without degrading the classifying accuracy. In the end, the recognition experiment of underwater acoustic images has been done and the result is satisfactory, which verifies the effect of this method.