Occupancy Tracking and Direction Detection using Machine Learning and Image Processing

Facing direction detection plays a critical role in human computer interaction (HCI), and has received significant attention in recent research, such as survei llanc e syst ems, comput er games, driving awarenes s recognition, assistive technologies and suspicious behavior detection. Currently the most popular approach for facing direction detection is optical-camera based facial feature extraction, such as such as skin color or eye gaze. The associated identification algorithms include a static head-pose estimation algorithm and a visual 3-D tracking algorithm for driver awareness monitoring. However, optical-camera based approaches raise concerns on privacy invasion and these devices do not usually work in a dark environment. And thus they are not an ideal option for many applications, such as rapidly increased smart home applications. With the popularity of smart home devices, unobtrusive HCI solutions are becoming more and more desirable, such as the use of passive infrared (PIR), and thermopile array sensors. PIR sensor is widely used as motion detector for automatic lighting and HVAC control, it has also been extended for occupancy moving direction detection.