Contour detection algorithm research based on the non-classical receptive field properties
|School||University of Electronic Science and Technology|
|Keywords||Non-classical receptive field Contour detection Gabor operator Cluster effect Inhibitory effect|
The edge of the image contains a lot of information and contour , edge detection and contour extraction plays an important role in computer vision, mathematical image processing and machine intelligence . So that the edge with the contour information is difficult to accurately extract , so a long time It is also related to the field of a challenge due to the complexity of the image background . The human visual system can easily identify the objects from the environment to extract the contour of the object and the edge information . Receptive field mechanisms play a significant role in this process to extract the edges and contours . The research and application of the receptive field properties can not only effectively improve the treatment effect , but also makes the processing results to match human visual characteristics . Traditional edge detection operator summarized the principles of representative operator , and on this basis, given the difference between the traditional edge detection and contour extraction . Human visual information processing mechanisms in the receptive field and the non-classical receptive field properties and their interaction mechanism , we propose a model contour detection method based on the non-classical receptive field toward selective inhibition . Based on the clustering characteristics of the primary visual cortex visual cells , this paper presents a contour extraction method based on the non-classical receptive fields of visual cells cluster effect . The two methods are in the natural image of the experimental results in the case of comparison with the traditional method , the proposed method can effectively extract the contour of the target object in a complex background , while suppressing and removing the interference of background texture . In order to make the results more persuasive , we use the method of quantitative determination of a variety of edge detection operator and the proposed method were compared . Quantitative comparison with the reference image , the two methods are proposed in this paper results are significantly better than the other algorithms .