The time people spend in front of computers has been increasing steadily due to the role computers play in modern society. Individuals who sit in front of computers for an extended period of time, specifically with improper postures may incur various health issues. In this work, individuals' behaviors in front of computers are studied using web cameras. By means of non-rigid face tracking system, data are analyzed to determine the 3D head pose, blink rate and yawn frequency of computer users. When combining these visual cues, a system of intelligent personal assistants for computer users is proposed.