In this paper, we design and construct a novel bio-heuristic character image classification algorithm based on current research of neurobiology and cognitive physiology, taking into account the biological learning mechanism of human brain, and using HMAX model and extreme learning machine. We focus on color image segmentation model based on the perception mechanism of visual cortex and a hierarchical color image perceptual segmentation model based on multi-feature is put forward. The algorithm effectively utilizes the spatial distribution of brightness, details and color space information of the image to segment the image for the first time, and uses corresponding model to integrate and select the segmentation results of multi-feature, to obtain the final segmentation results. The experiments show the advantages and practicability of the proposed bionic image target recognition algorithm through certain specific tasks. Simultaneously, it is also proved to have good recognition accuracy and fast learning speed.
展开▼
