Interactive Full-Body Motion Capture Using Infrared Sensor Networking
The article from Son Duong and Min-Hyung Choi from the University of Colorado Denver talks about in this journal the idea of traditional motion capture and the developments from the traditional rotascoping to the newer markerless traditional motion capture.
The paper presents new techniques of using multiple infrared devices that allow the process of information from multiple infrared cameras to allow enhanced data collection with the flexability and accuracy that is seen in most modern markerless motion capture presenting both the positives and negative parts of this new type of research.
Negatives:
Infrared allows for that collation of information in front of you; However, as you are using infrared, the particles or beams they use to collect information and reflect and bounce into other cameras faulting their main collation of data information. This means that the data can be skewered so cleaning is still mandatory at the end to ensure clean motion capture.
Your cameras can also be unsync'd. As infrared collects the information and presents in a 2D model presentation all information within it's 3D space that is captured, this is called the structure-from-motion approach (SfM). Although collation of information can be used to present a much clearer presentation of a 3D model using multiple sources, it's the placement of ensuring the same image and collecting that imagery at the same time to allow that triangulation that can be difficult.
Positive:
It allows the use of markerless tracking which means no plotting of information on the person itself and means that there is no necessity of time being wasted plotting on the character being used.
Conclusion:
The journal presented good feedback from the work and presented that the allowed for a bigger field of improvement and more reliable positioning for accurate tracking of the actor.
Comments
Post a Comment