Welcome to KineWheelArena - DLC’s documentation!

Kine Wheel Arena – DLC is an open-source experimental setup for markerless paw tracking of head-fixed mice running on a wheel. The apparatus is based on the open-source motion tracking system (KineMouse Wheel) designed by Richard Warren. It has been extended with a camera, lighting, and a controller to run both in synchronization. Tracking is done using a Deep Neural Network trained with DeepLabCut (DLC). The system is available for purchase at LABmaker.

Figure 1 – Video clip showing mouse paws annotated by Deep Neural Network.

Contents

• Requirements
  • Hardware
  • Software
    • KWA-Controller App
      • Usage
    • pylon Viewer
      • Install pylon Viewer
      • Open Camera
      • Load Camera Features
      • Recording Settings
      • Record
    • Anaconda/Miniconda
• Repository Contents
• Installation
• Usage
  • Workflow
  • (I) Record Videos
  • (II) Analyze and Label Videos
    • Create DLC Project for Inference
    • Inference using Deep Neural Network
      • (A) Run Inference on a Local Machine
      • (B) Run Inference on a Cloud Provider
    • Inference Results
• Example
• License