The Research On The Texture Image Recognition Based On The Spiking Neural Network

The third generation of neural networks is based on the Hodgkin-Huxley spiking neuron model.The functionalities of spiking neurons can be applied to deal with biological stimuli and explain complicated intelligent behaviors of the brain.Spiking neural networks always work with a large population of neurons.As a large-scale network of spiking neurons requires high computational resources to simulate,the integrate-and-fire neuron model and spike response model are usually regarded as a simplified Hodgkin-Huxley model.Since spiking neuron models are employed and information is encoded using the patterns of neural activities,learning mechanisms for spiking neural networks are very different from that in the first two-generations of classical neural networks.Deep learning is a class of machine learning techniques that exploit many layers of non-linear information processing for supervised or unsupervised feature extraction and transformation,and for pattern analysis and classification.The human vision system perceives scenes having variations of intensity and color,which form certain repeated patterns,called texture.Such characterizations of textures are generally applicable mainly to deterministic type of textures,such as,line arrays,checker boards,hexagonal tiling,etc.Texture is one of the significant characteristics used in identifying objects of interest or regions in an image.Texture analysis is important in many applications of computer image analysis for classification or segmentation of images based on local spatial variations of intensity or color.There are many texture analytical methods,such as structural analysis method,statistical method etc.In view of its multifaceted properties,texture analysis finds a lot of applications in many areas,e.g.,medical image analysis,automatic surface inspection,remote sensing etc.In this dissertation,firstly,considering the wavelet transform is widely used in image feature extraction and the similarity of the algorithm of wavelet transform and convolution neural network of deep learning,spiking neural networks with the ON/OFF neuron pathways are proposed to perform the feature extraction and the reconstruction for visual images.These constructed networks are based on the principles from the visual system,convolutional neural networks of deep learning and wavelet theory.By this way we try to simulate how the human brain uses volition-controlled method to extracts useful image information.Furthermore,we decompose each texture sample with the established networks and calculate the normalized energy of the obtained sub-images at different scales.These energy values are used as features for texture classification.The simulation results show that the spiking neural network can extract the main information of images so that the images can be accurately classified using the main information.In the other hand,the Grey-Level Co-occurrence Matrix(GLCM)algorithm is widely used in visual images for texture feature extraction,image structure characterization analysis and texture classification and segmentation.In this dissertation,a spiking neural network has been presented inspired by the deep learning algorithm and the biological mechanisms in the primate brain,which has excellent performance in terms of image feature extraction.The improved GLCM algorithm is used to train this spiking neural network and simulate the brain's ability about extract key information.These extracted feature information will be utilized to classify different texture image and achieve the texture segmentation.Experimental results in this article show that this combination of the GLCM and spiking neural network can effectively extract the image co-occurrence features,and also obtain satisfactory classification and segmentation effect.In this article,we try to achieve button recognition in a specific category through the designed experimental method.In addition,a portable guiding device has been developed,combined with the proposed algorithm,it can provide effective information support for the blind.