November 13th, 2013 at 5:00pm PDT, Agenda and Minutes
Dick Puk, Myeong Won Lee and William Glascoe attended. Don Brutzman joined 20 minutes into the call.
BVH to X3D Interpolators
Euler to quaternion conversion
Myeong sent an algorithm to convert Euler angles to a quaternion to Don. There are several ways to convert Euler angles to quaternions, Don said. The conversion can be offline because it is not part of an X3D browser. Don reminded us this conversion is an implementation issue only.
BVH to LOA1, 2 and 3. Need mapping from BVH channel hierarchy to each LOA. Myeong is programming an application to convert BVH to the appropriate LOA.
LOA 1, 2 and 3
Definition and naming
Joe and Don are resolving joint names between the different LOAs.
Joint name: feature point, and landmark
Applications for LOA 1, 2 and 3
What is the best application for each level? Myeong thinks the medical applications warrants higher LOAs. William suggests the Use Case will dictate the LOA required. Dick agrees with William assertion. Don suggests we could had intermediate levels to map to some the Use Cases that are emerging in the market place. We are working on an intermediate LOA for hands, feet and facial animation. Myeong thinks we should make examples for LOA 2 and 3. Dick suggest we experiment with the same motion but use two LOAs. Do you always use a superset or allow additions to LOAs. Dick asserts it is a difference between specification and usage. How do you reconcile a situation of using a LOA3 (i.e., full up body ) but only animate a few of the joints? Dick wants to change the labels (title of Don's figure [HAnimHigherLoa-draft4.png]) specify level of articulation. Each level of articulation defines a "full" skeleton but higher levels resolve more joints (and segments or bones).
Scheduling next meeting
We will meet the 1st Wednesday of December, the 4th at 5:00pm PST (5th, 10:00am KST)