Understanding big complex networks is an important task in many real applications. Graphs are used for representing complex networks. Graph summarization by reducing the size of a graph and the complexity of graph structure makes massive network data more understandable. In this paper, we propose a simple and efficient graph summarization method, called VGS, that exploits visual graph mining techniques. VGS graph summarization method is based on conventional machine learning using graph structures which avoids calculation of pairwise similarities. VGS identifies nodes with high degrees forming dense regions and nodes with low degree forming sparse parts of graph. First central nodes with degrees higher than the surrounding nodes are identified. Connected central nodes are joined into super nodes. And connected neighbor nodes are joined in the new super node. The greedy summarization algorithm VGS can reach the summarization ratio from 27% to 50%. CCS CONCEPTS: Network, Graph summarization