A Novel Bio-inspired Image Recognition Network with Extreme Learning Machine

Abstract

This paper presents a novel bio-inspired network for image recognition. The HMAX model and the extreme learning machine (ELM) are combined, to construct a five-layer feed-forward network: S1-C1-S2-C2-H. The previous four layers, originating from HMAX, provide robust feature representation of specific object, and the feature classification stage at the H layer is implemented with ELM. The HMAX model simulates the hierarchical processing mechanism in primate visual cortex, to calculate feature representation. As a biological learning algorithm for SLFNs, ELM learns much faster with good generalization, and performs well in classification applications. Our experimental results show effective accuracy performance with fast learning speed.