Research On Interactive Image Segmentation Algorithm Based On Improved Combination Of Superpixel And Graph Cut

Image segmentation is one of the hotspots in the field of computer vision,there is a very important application in image processing and other related fields.It is an important part of image engineering,which is widely used in computer vision,pattern recognition and other fields.Color image segmentation use color,texture and regional characteristics to extract the target region from the background of the image,so as to obtain the human interest.In the last ten years,many excellent algorithms have appeared in the field of image segmentation.The graph cut is a segmentation algorithm based on graph theory,it dispose the pixel value of image mainly by constructing an energy function,by minimizing the energy function to get optimal solution,this method can obtain the global optimum,it has better robustness and higher segmentation precision.However when it deal with big and high resolution image,it will cost too much time.This affects the application of graph cut algorithm in reality.Super pixel is consisted by many similar characteristics' pixel,there are many similar characteristics in same pixel blocks,such as brightness and color.And the pixel blocks can preserve texture features and edge information of the image,the boundary can fit the contour of object segmentation to a certain extent,pretreatment is acted on before formal segmentation of the image,so that it reduce the complexity of segmentation and improve the segmentation efficiency.It can improve he efficiency of image segmentation when combin super pixels and graph cut,especially for the need for iterative algorithm which can reflect more efficient,especially for iterative algorithm,due to the use of super pixels instead of pixel value to segment,it will lead to a decline in the accuracy of segmentation,and when the number is too little,the results will appear serious deterioration.To overcome the phenomenon of deterioration in segmentation which is based on super pixels in the image segmentation algorithm of Grabcut when the number of super pixels is low.It put forward a segmentation algorithm which is combined with bayesian classification and SLIC((Simple Linear Iterative Clustering)to improve Grabcut.Firstly,to use the SLIC algorithm for image clustering,after that it use RGBmean value of evey pixel blocks as the pixel dot to form contacted Graph cut model,and then use bayesian classification to classify pixels in the model,it apply SLIC to classify the pixels in the second time.In order to estimate the value of GMM,this algorithm use the mean of the super pixels color value to represent the all pixels color value.Finally it used min-cut algorithm to get the optimal segmentation of graph.The experimental results show that the algorithm reduce segmentation mistake and improve efficiency of segmentation.