A Computer Vision-Based Approach for Biological Phenotyping

Brief Description: This research computationally analyzes the behavior patterns of animals, providing insight to phenotypical behavior

Participating Labs: Smith Lab, Fuse Research Lab, Moffatt Lab, Graphics and Image Analysis Research Group

Research Purpose: In this research, we propose applying color and motion based tracking to study foraging behavior in ants, odor sense response in the mouse, ecdysis in the hornworm. These significant complex phenotypes have thus far eluded automated approaches for quantitative analysis. We present a vision-based algorithm for robust tracking that can be easily applicable to multiple organisms. The experimental results demonstrate the accuracy of tracking and phenotype determination under conditions of complex body movement, partial occlusions, and body deformations.

Publications

Software

• In Progress

Researchers

• Smith Lab
  • Chris Smith, Principal Investigator
  • Lawrence, Undergraduate researcher
  • Jennifer Placek, Graduate researcher
• Fuse Research Lab
  • Megumi Fuse, Principal Investigator
• Moffatt Lab
  • Chris Moffatt, Principal Investigator
  • Emily Merchasin, Graduate Researcher
• Graphics and Image Analysis Lab
  • Ilmi Yoon, Principal Investigator
  • Philip Burkhard, Undergraduate Researcher
  • Jennifer Lee, Undergraduate Researcher
  • HendraLim^?, Graduate Researcher (graduated)
  • Alan Shimoide, Graduate Researcher (graduated)
• Center for Computing for Life Sciences
  • Mike Wong, Staff Researcher