Alert icon
We're changing our privacy policy. This stuff matters.  Learn more  Dismiss

Style-based Inverse Kinematics

Loading...

Sign in or sign up now!
Alert icon
Upgrade to the latest Flash Player for improved playback performance. Upgrade now or more info.
5,787
Loading...
Alert icon
Sign in or sign up now!
Alert icon

Uploaded by on Mar 28, 2008

Keith Grochow, Steven L. Martin, Aaron Hertzmann, Zoran Popović

We present an inverse kinematics system based on a learned model of human poses. Given a set of constraints, our system can produce the most likely pose satisfying those constraints, in realtime. Training the model on different input data leads to different styles of IK. The model is represented as a probability distribution over the space of all possible poses. This means that our IK system can generate any pose, but prefers poses that are most similar to the space of poses in the training data. We represent the probability with a novel model called a Scaled Gaussian Process Latent Variable Model. The parameters of the model are all learned automatically; no manual tuning is required for the learning component of the system. We additionally describe a novel procedure for interpolating between styles.

Our style-based IK can replace conventional IK, wherever it is used in computer animation and computer vision. We demonstrate our system in the context of a number of applications: interactive character posing, trajectory keyframing, real-time motion capture with missing markers, and posing from a 2D image.

Category:

Science & Technology

Tags:

License:

Standard YouTube License

  • likes, 0 dislikes

Link to this comment:

Share to:
see all

All Comments (7)

Sign In or Sign Up now to post a comment!
  • Is the 2d probability map how things are actually handled internally or just a simplification for the sake of demonstration or easiness of direct manipulation by a human or somthing?

    And does it take time in consideration, i mean, does the pose in the previous N frames affect which pose the system thinks is the most probable for the current one?

  • Very clever research! Is the circular arrangement of data points at 2:45 due to the cyclic nature of walking? Are those 2D plots generated by a Kohonen self-organizing map?

  • This is beyond me. I recently developed 2D skeletal animation with inverse kinematics, but it's so much easier in 2D. I'm a hobbyist, though.

  • wow, thats alot of math! lol

  • I am very interested in your research, do u have papers published on it?

  • Dude, CliffBurton, you're a dumbass.

Loading...

Alert icon
0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more