Adding Cartoon-Like Motion to Realistic Animations No Comment

Project Type: Master’s Thesis
Supervisor: Dr. Chris Joslin
Project Components: Motion Capture, Computer Programming (Python, Autodesk Maya), User Studies, Technical writing


Nancy and Michelle in their respective motion capture sessions.


In comparison to traditional animation techniques, motion capture allows animators to obtain a large amount of realistic data in little time. The motivation behind our research is to try to fill the gap that separates realistic motion from cartoon animation. We want to allow classical animators to produce high quality non-realistic video games content and animation films in a shorter amount of time.

Curve Editing and User Interface

Users edit the animations by manipulating a curve that determines how much to modify the animation by. We created three main editing categories: angular exaggeration, jump trajectory exaggeration and animation morphing (neutral to stylized-feminine walk).

Result Samples


Jump Trajectory Exaggeration

Neutral Walk to Stylized Cartoon Feminine Walk


Data Collection and User Studies

  • Total of 15 participants recorded various movements through our motion capture sessions (~2 hour/session)
  • Total of 30 participants across two user studies  (~1 hour/session)
  • Both user study results suggest that the user interface is easy to use and intuitive for amateur users
  • Both user study results show that the animation’s realism was successfully reduces.
  • User study results are not conclusive on whether our algorithm appropriately increases visual appeal.

Research Contributions

  1. A simple method that changes motion by modifying animation curves.
  2. A simple curve interface as a means of interaction with the system.
  3. Two user studies to evaluate the functionality and usability of the system.



  • System successfully reduces realism and is easy to use by novice users.
  • Curve based interface illustrates the animation changes appropriately.


  • Our algorithm does not function very well at extreme data points (large slowdowns, large exaggerations) and can produce very unappealing results when the editing coefficients are too large.
  • Feminine walk coefficients could be modified to reduce the unstable look of certain results

Leave a Reply