利用非经典感受野竞争机制实现有效图像表征的方法

摘要

In this paper,the physiological mechanism of non-classical receptive field is utilized and a hierarchical network model for image representation is designed based on neurobiology.It is different from the contour detection,edge detection,and oth?er practices using the classical receptive fields,it simulates the non-classical receptive field's physiological mechanism which can be dynamically adjusted according to stimulation for image local segmentation and compression based on image neighborhood region similarity,thus to realize the inner image representation in neural representation level and convenient for extract the semantic fur?ther.%论文是在人类视觉的基础上,利用神经节细胞非经典感受野的神经机制原型对图像实现通用表征.该模型利用生物视觉处理中的反馈机制设计了一种神经计算回路,模拟了非经典感受野在不同的刺激状态下能够动态调节的生理特性来对图像进行基于邻域相似度的图像局部分割与压缩表征,从而在神经表征层面实现了图像的内在表征,以便于进一步提取图像的语义,实现图像的理解.为更加有效、充分地表征图像提供了一种新的研究思路.