Description

Superpixels are becoming increasingly popular for use in computer vision applications. Image segmentation is the process of partitioning a digital image into multiple segments (known as super pixels). In this paper, they have proposed the methodology called as Lazy Random Walk. In our work we are proposing the methods called as Patches and labels. This techniques performs based on the color in the query image. Hence the image can be segmented into multiple parts by its colors. In this work we present Optimized superpixel segmentation Using our proposed system Called Patches and labels. In previous they have developed superpixel segmentation using Lazy random walk methodology. This method optimized from Random walk technique. By reducing the speed of the Random walk process the achieved this Lazy random walk. In our work, the Patches and label methods are based on Clustering and Color mapping. Initially the image is preprocessed by filtering using median filter for noise removal. Then the processed image mapped by its color in 9 levels. 9 levels of color mapping provides the mapped regions and colors, from that the process clustering occurs to grouping of pixels. This provides view of object with light background. Using these techniques we segmented the query image. Through the Probability map we are identifying the area to be segmented. The overall process segments the object from the background image. Which And then by the segmented result we are getting the segmented value. By using the result values we are Using our process to the Classification purpose. In this classification method we are classifying the query image.