Abstract
Key point detection is still a challenging issue in pattern recognition. With the recent developments on complex network theory, pattern recognition techniques based on graphs have improved considerably. Key point detection can be approached by community identification in directed complex network because image is related with network model. This paper presents a complex network approach for key point detection in video monitoring image, which is both accurate and fast. We evaluate our method for square and subway station video monitoring images. Results show that our algorithm can outperform other traditional method both in accuracy and processing times.