With the development of the Internet, more and more people actively interact with others via online social networks. Potentially, people can hide themselves in the dark and continually gather information from other users from the Internet. To assist individual users to protect their privacy and security, in this study a computational approach for abnormal attention detection will be presented. The proposed approach can detect abnormal attention from the local view of a user, without invading other people’s privacy. We then move on to focus on the online interpersonal surveillance which is an excessive, unreciprocated and persistent attention. We address the issue of interpersonal surveillance by asking the question, “who is surveiling you through social networking?”. This is a challenging question, as interpersonal surveillance is a victim-defined behaviour and often occurs without a visible trace. Viewing a network as interconnected agents who interact through posting and reading information, we provide a measure to quantify the level of attention a person pays towards another from a global view. This measure allows us to capture online interpersonal surveillance.Through theoretical and experimental analyses, we show that our method distinguishes and detects online interpersonal surveillance.