Animal behavior Analysis in an Aerosol Exposure Chamber

Apr 5, 2024

Animal Behavior Analysis in an Aerosol Exposure Chamber

Project Description

  • Motivation: Investigate whether aerosol exposure affects animal’s behavior and walking patterns.
  • Analyzed the behavior of two lizards in different conditions:
    • One in a non-aerosol exposure chamber.
    • One in an aerosol exposure chamber.

Contribution

Tracking

  • Methodology: Tracked lizards using a deep learning network trained with a self-labeled dataset.
    • Model: YOLOv8
    • Dataset: Generated 1,701 labeled images from recorded videos, manually labeling each lizard joint.
      screen reader text
      Figure 1: Lizards (Yellow and Black Dot) Tracking
      screen reader text
      Figure 2: Black Dot Lizard Tracking in a Non-aerosol Exposure Chamber

Analysis

  • Indices: Analyzed lizard behavior using 4 indices:
    • Stride Speed
    • Movement
    • Speed
    • Leg Angle
  • Visualization: Presented the analyzed data using Python.

screen reader text
Figure 3: Analysis of Lizard Stride Speed
screen reader text
Figure 4: Analysis of Lizard Movement
screen reader text
Figure 5: Analysis of Lizard Speed
screen reader text
Figure 6: Analysis of Lizard Leg Angle

Skills

  • Python: Analysis, Visualization, Deep Learning Training and Inference
  • PyTorch: Deep Learning Training and Inference
  • OpenCV: Video Streaming, Frame Extraction
Computer Vision Artificial Intelligence Data Management Research
Hyungjun Doh
Hyungjun Doh
Master’s Student

My research interests are Human-AI interaction and its practical applications, with a specific focus on Extended Reality, Task Guidance Systems, and AI-infused interfaces.