3D Model Viewing from User’s Point of View

Using Depth Camera and Head Tracking

The 3d model viewing system we developed allows the user to view 3D objects from different points of view using depth camera and head tracking.

This project offers a solution using a facial recognition algorithm provided by the Intel Perceptual Computing SDK a depth image provided by the CREATIVE camera. The system combines the two images into one image that includes the distance of each pixel in addition to its RGB value. From this information the system figures the user’s face coordinate in 3D space in order to understand to user’s view point on the screen (object).
This solution lets the system to transform the object projected on the screen according to the user’s view point. In other words, the solution allows the user to observe the object as it should look from many different points of view. This effect is achieved using 3D perspective projection transformations on a frustum 3D space.
The projects implements a GUI to use the system on a large variety of 3D models and allows the user to select the sensitivity levels of the system’s response to his movements in space.