Abstract: Need of image retrieval has increased with the addition of digital images. Construction of graph for capturingperceptual grouping cues of an image is important for graph based image segmentation. Basic CBIR system captureonly the local visual feature, thus to cover the gap between local feature and high level visual perception we proposedthis system. This technique introduces graph-based image segmentation method in order to retrieve the content-basedimages. In this CBIR system first, query image over-segmented into multiple super-pixels then extraction of visualfeatures e.g. color, shape and texture. Affinity graph construction is there for building a graph from super-pixels. Finallygraph edit distance will be used for graph matching between query image and database images. Then images withsmallest edit cost will be retrieve and display relevant result to user. Thus proposed CBIR system using graph basedimage segmentation will improve performance and accuracy.

Keywords: Content Based Image Retrieval, Image Segmentation,Graph Construction,Graph Matching,Super-pixel.