High-Resolution, Real-Time-Geometry Video Acquisition System

A geometry video acquisition system that measures both geometries and positions of objects accurately. The data acquisition speed is 90fps and 266,000 points per frame. Acquisition, reconstruction, and display are simultaneously realized at 30fps.

Enhanced Life
This technique takes full advantage of a digital-light-processing (DLP) projector and uses a unique method to synchronize a camera with the projector. The phase-shifting algorithm, which has unique advantages over other algorithms, is used to its extreme. It uses the minimum number of fringe patterns as it obtains the maximum possible information.

Goals
To measure absolute motions of both geometric shape and positions of deformable objects in real time to generate geometry videos and to find potential applications in various fields.

Innovations
The main innovation is integration of the whole system, to develop highly efficient and robust software algorithms, and to measure absolute coordinates of objects pixel by pixel in real time.

Vision
Capturing 3D geometric data can be as simple as acquiring 2D images. Recording 3D videos is no longer a difficult task. Without knowing the internal technology, you can own one device and capture your own geometric data immediately.

The geometric video acquisition system can be applied in many fields. One immediate application is 3D modeling of facial expressions in games, for example. Since acquisition, reconstruction, and display are done in real time, digitizing players and putting their faces into virtual game worlds will be feasible in the near future. Another application is animation and movies, where real captured video data can be used in animation immediately. The system can also be applied to plastic surgery and homeland security. The geometry of the object overcomes many shortcomings of security checking techniques based on 2D images, such as facial recognition, finger prints, etc. Other applications include motion capture, fast prototyping, virtual reality, feature films, dentistry, human-computer interaction, robotic vision, reverse engineering, industrial measurement, and inspection.

Emerging Technologies Sketch

Contact
Song Zhang
Harvard University
Szhang (at) fas.harvard.edu

Contributor
Dale Royer
Geometric Informatics Inc.

Shing-Tung Yau
Harvard University