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

• Main Author: Wu, Jason Weixiang.
• Other Authors: Cham Tat Jen
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/48542
• Institution: Nanyang Technological University

Child minding has always been a challenging task for most parents especially for those with kids aged one to five years old. This is the crucial period of growing up whereby children begin to walk and run about, start to play with their toys and the time when their curiosity about the surroundings slowly develops. During this period, children often unknowingly place themselves in danger which makes looking after them a tedious task for caregivers. The proposed automated surveillance system in this project aims to enable caregivers to minimize the amount of attention needed and at the same time provide alerts and ample time for caregivers to respond accordingly to the situation. Firstly, the proposed system will require the user to identify the estimated location of the child and define a danger zone of which the child may endanger himself if he should come into contact with. After which, background subtraction was used to segment the moving and non-moving parts of the surveillance video. The system will then use kernel-based mean shift tracking using the Epanechnikov kernel to track the child and prompt alerts if the child comes into contact with the danger zone. In this manner, the child-minder will then be able to carry on with their routine tasks without requiring additional attention. The system was tested using surveillance videos recorded in various real home environments where children under the age of five resided. The system could effectively track the children and provide audible alerts when they approached dangerous areas. This shows that the system can provide caregivers with a means to reduce the amount of attention spent on monitoring children. Various possible built-in features to improve the system such as a task scheduler and the collection of statistics of the child’s daily activities were also discussed in the later part of the project.