Specialized Mappings Architecture

Pose Estimation Examples

 


 

3D hand pose estimation examples

 

Example mean-output algorithm (videos)

 

Mean-output algorithm applied to sequences of segmented hands (hand palm view)

Input segmentations are shown on the right

MO estimates are shown in upper-left corner

A faster version

 

Mean-output algorithm applied to sequences of segmented hands (unrestricted view)

Input segmentations are shown on the right

MO estimates are shown in upper-left corner (after small delay)

 

Example mean-output algorithm (still images)

 

Abbreviations

RV=Real Video

MO=Mean-output result

MS=Multiple-sampling

S1=Best sample

S2=Second best sample

...

 

 

Hand palm views

Unrestricted views

 

Example multiple-sampling algorithm (still images)

 

Hand palm views

Unrestricted views

 

2D body pose estimation examples

 

Example mean-output algorithm applied to a sequence of segmented human bodies

 

 

For more examples/details see:

 

Rómer Rosales and Stan Sclaroff
The Specialized Mappings Architecture
(pdf|ps)

Submitted (03/03).

 

Rómer Rosales, Vassilis Athitsos, and Stan Sclaroff

3D Hand Pose Reconstruction Using Specialized Mappings

In Proc. IEEE International Conference on Computer Vision (ICCV). Presented at ICCV, Vancouver, BC, Canada, Jul 2001.

 

Rómer Rosales and Stan Sclaroff

Learning Body Pose Via Specialized Maps

In Proc. Neural Information Processing Systems NIPS-14, 2002. Presented at NIPS, Vancouver, BC, Dec 2001