Visual Information Encoding Modeling From the Perspective of Cell Assemblies

Abstract

This paper aims to explore the visual information encoding mechanism based on neuronal cell assembly theory, focusing on improving the interpretability and credibility of visual information processing models. By comprehensively applying multidisciplinary theories and methods such as neuroscience, computational neuroscience, and information theory, the effect of neuronal activity in the process of visual information processing is analyzed step by step, and the way visual information is encoded in brain is revealed. It is worth emphasizing that results in this paper is just based on the depth of understanding of neuroscience today. Much unknown remains mysterious so only conjectures and assumptions can be made. It is found that the neurons classify, extract, and reconstruct scenes of complex visual features from the acquired images. And visual modeling idea based on Hebb’s theory about cell assemblies is similar to the real physiological processing process of visual information, and has great potential to solve real-life research problems.