In order to assist people’s everyday life more precisely, intelligent monitoring is necessary to reduce human efforts on watching monitor screens. For the changing
environment, making decisions only on target recognition by fixed model is not sufficient at all, which is difficult to increment the knowledge and also neglect the implicit valuable information hidden in association between people and the surrounding objects. In this paper, inspired by human interactive learning model, we propose a human-interactive intelligent monitoring system which can autonomously learn people’s habits according to the human-object-time association rules day by day, guided by human’s expert knowledge at childhood stage or while getting in confused. Depends on its observations and learnt knowledge, while some anomalies happen, give real-time response. Without any prior knowledge in advanced, make it
keep learning after observations from user’s environment, the precise and customizable services can be provided to every different user. Due to the advantages of Fog Computing, multitier fog structure is involved to deal with the analytics and transmission of the huge volume of data to achieve real-time responses.