Assisted child-minding based on visual activity monitoring in a home camera surveillance system

Abstract

Monitoring a child can be demanding these days. This is especially so in our current society where both parents are busy with work both inside and outside of the house. Children might get injured when the parents are not paying enough attention. To deal with this problem, an assisted child-minder system is being proposed. The aim of this project is to monitor the child through visualization methods with a given dataset that is obtained through surveillance cameras. Using this system, the user will be able to track the movements of the child. Also, a boundary can be defined as such when the child crosses the boundary, the parents will be notified. This system is developed in visual studio using C++ and the OpenCV library. Firstly, different background subtraction techniques were implemented to separate the dataset’s foreground and background. Secondly, mean shift algorithm was implemented to track the child’s movement. Lastly, a boundary is defined to detect if the child entered the restricted zone. Although the mean shift algorithm is successful in tracking the child in dataset 1, it is not very successful in dataset 2. There is definitely still room for improvement to increase the algorithm’s accuracy.