A study of children facial recognition for privacy in smart TV

Abstract

Nowadays Smart TV is becoming very popular in many families. Smart TV provides computing and connectivity capabilities with access to online services, such as video on demand, online games, and even sports and healthcare activities. For example, Google Smart TV, which is based on Google Android, integrates into the users’ daily physical activities through its ability to extract and access context information dependent on the surrounding environment and to react accordingly via built-in camera and sensors. Without a viable privacy protection system in place, however, the expanding use of Smart TV can lead to privacy violations through tracking and user profiling by broadcasters and others. This becomes of particular concern when underage users such as children who may not fully understand the concept of privacy are involved in using the Smart TV services. In this study, we consider digital imaging and ways to identify and properly tag pictures of children in order to prevent unwanted disclosure of personal information. We have conducted a preliminary experiment on the effectiveness of facial recognition technology in Smart TV where experimental recognition of child face presence in feedback image streams is conducted through the Microsoft’s Face Application Programming Interface.