Accelerated HMAX model based on improved SIFT feature points

Object recognition technology is an important research field of image understanding and computer vision, with its wide range of application, it attracts more and more attention. HMAX was proposed as a simple and biologically feasible model for object recognition, based on how the visual cortex processes information. However, computational cost is the biggest obstacle of this model. This paper aims to improve HMAX, and the work of this paper is as follow: 1. By studying the directional characteristics of Gabor filters, a convolution layer sparsing method is proposed to reduce the time-consuming of convolution layer. 2. By studying the extracting techniques of feature point, a new SIFT feature extraction algorithm is proposed to resolve the redundancy of patches in sampling layer. At the end of this paper, we apply the improved HMAX models to Caltech101 database. By comparing with the original model, the experimental results show that improved HMAX has a better performance.