The Angle Information Extraction Mechanism Based On The Primary Vision System

Corner is an important dominant point in contour features,which can not only preserve image features to the maximum extent,but also effectively reduce the amount of information.Corner recognition is one of the important bases in the field of image recognition.The traditional industrial algorithm for corner recognition is based on the mathematical method,which can identify corner points by the similarities and differences of various numerical values on the corner points and the pixel points in the area near the corner points.However,there are some limitations.With the development of The Times,people's demand for image recognition technology is increasing day by day.It requires not only accurate identification but also fast and low energy consumption.After the beginning of the wave of brain research in various countries,image recognition technology based on human visual system has become a new research direction.The human visual system has the characteristics of rapid,accurate,low energy consumption,low space demand and so on.At the same time,as the most sophisticated organ of human beings,the human brain is still very simple to study.It is of great significance in medicine and physiology to reveal the operation mechanism of human brain.In order to study the mechanism of Angle information extraction based on human visual system,the concept of Angle should be defined first.At present,the definition of corner has not been determined.There are two main theories,one is to speculate the position of Angle vertex based on the information of adjacent edge of image.This theory does not make full use of the region information around the diagonal,and the position information of the calculated Angle vertex is consistent with the real position of the Angle vertex,which is too accurate and does not accord with the common sense of human vision.This paper proposes a corner information extraction mechanism based on the primary vision system,which can obtain the corner vertex position information and other corner related information without relying on the image edge position information.By using the differential cell receptive field constructed by simple cells,the information of the region around the corner is collected in a two-dimensional discrete distribution mode,and the Angle vertex position and other related information are calculated accordingly.The position of the Angle vertex calculated by this method will be offset to the Angle opening direction,and the deviation of the position and the offset have a certain relationship with the degree of the Angle.In addition,it is found in the experiment that the detection effect varies greatly when different resolutions are used to process the same picture.Therefore,it is speculated that the resolution selection of human visual system may be related to the size of the processed image.The phenomenon of false vision is a special phenomenon of human visual system,which can reflect the characteristics of human visual system to a certain extent.In this paper,the position deviation of corner points and the resolution selection of human visual system under the mechanism of corner information extraction are discussed in combination with the graphic induction experiment of muller-lyre vision error.The mechanism of Angle information extraction and the possibility of human visual system resolution selection are illustrated.