Pixel Vision transforms the Microsoft Kinect into an electronic “eye” that passively detects and records the human form as a dimensional point-cloud and body-tracking tool.

Using real-time range image information in place of traditional full-color video, a variety of 3D image data can be captured, separated and massaged in ways that allows the human form to be mapped and used for multiple applications. By placing “hacked” Kinects in many locations, Pixel Vision becomes a connected, passive, low-cost system for gathering an influx of human body and movement data. Pixel Vision detects multiple human forms moving through space. With skeleton tracking, surveillance of human movement becomes even more detailed as the cameras are able to identify and isolate limbs, body parts, certain joints, and even facial expressions and recognition. This gives us the power to isolate and study specific motions and body parts that could become useful tools in medical research, health monitoring, safety & security, and interactive media applications.

The Microsoft Kinect camera system is able to obtain the depth image information to form a mesh of pixels that represent the captured human form. Code is written into the software to isolate the body and remove the background environment. Using this technology the depth information can be abstracted and down-sampled to give a dot array and used as a potential device to map human movement and record the x,y,z coordinates.

Multiple Kinects can be combined to form a more complete 3D image through triangulation. The more complete form becomes useful for body-part identification. In the case of human scanning, skeleton tracking can be used to identify the limbs, torso certain joints, and even facial recognition. This gives us the power to isolate and study specific motions and body parts that could become useful tools in medical research and security.