Novel method for Image Classification using Graph approach

Computer Engg.

Graphs are important in modeling complicated structures, such as natural scenes, animals, and flowers, images in our daily life, social networks and the web. Data representations in graphical forms are an important research topic due to the suitability of this kind of data structure to model entities and the complex relations among them. In order to classify images we are proposing to combine a powerful graph-based image representation and frequent approximate sub-graph (FAS) mining algorithms. From the many approaches for image classification, graph based approach is gaining popularity due to its ability in reflecting global image properties. In this paper, VEAM (Vertex and Edge Approximate graph Miner) algorithm is used for mining frequent connected subgraphs over undirected and labeled graph collections. Some Slight variations of the data, are allowed in this algorithm, by keeping the topology of the graphs. And, Graph-based image representation is by using dynamic region merging technique can tolerate some differences for grouping meaningful regions in an image.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.