Research On Algorithm Of Color Image Segmentation Based On PCNN

Image segmentation has a main function is to preprocess in image recognition and computer vision research,and it is the key step in image processing.The purpose of image segmentation is to divide the image into multiple non overlapping parts.At present,the common segmentation algorithm is roughly divided into two types: the traditional image segmentation algorithm and the image segmentation algorithm that combined with other related theories.PCNN(Pulse Coupled Neural Network)is a new generation of artificial neural network,which simulated the biological vision.The segmentation algorithm based on PCNN has gotten more and more attention from domestic and foreign scholars,because it has high accuracy and wide adaptability performance.And the development of multimedia technology and the popularity of various visual applications,make people pay more attention to color image processing.In recent years,researchers gradually pay more attention to co lor image segmentation.Therefore,the research of color image segmentation based on PCNN is particularly important.This paper firstly describes and analyzes the basic working principle and characteristics of the PCNN model and typical segmentation algorithms.Then,through these studies to improve the PCNN model.When PCNN is used for image segmentation,there are many problems.For example,it has many parameters that is difficult to set,and the image segmentation results are sensitive to the parameters.Through the maximum entropy criterion,an adaptive image segmentation algorithm is designed.It is based on global threshold obtained by OTSU and dynamic threshold obtained by improvement of the PCNN model.Experiments show that the designed algorithm can achieve the PCNN automatic image segmentation and has good segmentation results.This algorithm reduces the manual intervention and has a certain universality.At this stage,researches on color image segmentation based on PCNN is less,and has some problems.In view of the deficiency and defect of color image segmentation,this paper analyzes the color space model of color image,and find the suitable color space model to decompose the color image data.Combined with the characteristics of the regional growth method and PCNN model,through the in-depth analysis of these two methods,this paper use the PCNN model to simulate the regional growth method and make it suitable for color image segmentation.The improved PCNN model is used for each component of t he color space model.The final segmentation results are obtained by using the combination of the probability criterion to combine each component segmentation results.By writing the algorithm code,the algorithm is verified by the simulation,and the performance of the algorithm is evaluated using objective indicators.Experimental simulation results show that t his algorithm can achieve accurate segmentation results for color images,and has good generality.