[TC]2 reports that the powerful feature of the HAE is that it is based on a statistical sampling of a large number of previously scanned human subjects. Because it does not require the user to be scanned, it can be accessed anywhere.
The pool of 3D scan data comes from [TC]2's SizeUSA human sizing project of the U.S. population, which contains 3D body scans of more than 10,800 individuals, including age groups ranging from ages 18-65+ in all heights, weights and ethnicities.
Using this data and extracting more than 3,000 reference points from each body scan, it is possible to create a very realistic avatar of a person using a very small number of inputs using a statistical technique called principal component analysis.
An example set of input data points a person would use to create his or her avatar are items such as height, weight and age demographics, as well as a few body measurements such as chest, waist, seat circumference and inseam measure. In total, only a dozen responses are required to very closely replicate the real person, [TC]2 reports.
A real test of HAE's capabilities is to take a human subject who is not part of the statistical reference pool, create an avatar with the Human Avatar Engine, and then compare that with a full 3D body scan of that individual.
Unlike parametric computer generated human models, the output of the Human Avatar Engine is much more realistic and human-like and less "cartoon-ish," [TC]2 reports.
The HAE can be tuned to generate avatars in any mesh density and can include internal joint rigging locations for animation, allowing its easy accommodation in any existing virtual world or software application that utilizes human models.
for more information: www.tc2.com