DeepLabCut Workshop

A one day hands-on workshop on the use of the famous DeepLabCut software to analyze videos of animal behavior, led by Drs. Mackenzie and Alexander Mathis. Participants will be helped to pre-install the DeepLabCut software on their laptop and can bring their own data for a first experience in analyzing it.

Date: July 1 2025, 9:00 - 18:00.
         *Reception desk will open at 8:30AM (Access to the B250 Seminar Room).

Lecturers: Mackenzie Mathis (EPFL, Switzerland) and Alexander Mathis (EPFL, Switzerland)
Assistants: Aleksandra Gavrilova (OIST) and Saffira Yan Tjon (OIST)

Location: Sydney Brenner Lecture Theater (Seminar Room B250), OIST

Language: English

Registration fee: 

• OCNC2025 participants: free
• OIST personnel and students: free, registration required (no lunch provided)
• non-OIST students: 3000 JPY, includes lunch and coffee breaks
• non-OIST scientists: 6000 JPY, includes lunch and coffee breaks

Workshop Schedule

9:00 - 10:00 | Introduction to Computer Vision & DeepLabCut

   ●    Fundamentals of machine learning for behavioral analysis
   ●    Computer vision principles and pose estimation
   ●    DeepLabCut framework overview and capabilities

10:00 - 10:30 | Hands-on: Creating Your First DeepLabCut (DLC) Model

   ●    Setting up a DLC project
   ●    Strategic keypoint selection

10:30 - 11:00 | ☕ Coffee Break

11:00 - 12:30 | Hands-on: Labeling Training Data and kicking off training

   ●    Efficient labeling workflows
   ●    Quality control in annotations
   ●    Best practices for robust training datasets
   ●    Prepare datasets for training

12:30 - 13:30 | 🍽 Lunch Break

Models train during lunch break

13:30 - 14:30 | Model Evaluation & Inference

   ●    Model evaluation (metrics and results)
   ●    Cross-validation and performance assessment
   ●    Live tracking capabilities

14:30 - 15:30 | Kinematic & Behavioral Analysis

   ●    Lecture & Discussion: Methods for behavioral analysis and joint modeling of neural and behavioral data

15:30 - 16:00 | ☕ Coffee Break

16:00 - 17:30 | Practical: Behavioral analysis
   ●    Kinematic analysis of pose data
   ●    Action segmentation and behavioral clustering
   ●    Joint modeling of neural and behavioral data
   ●    Working with example datasets

17:30 - 18:00 | Advanced Topics & Wrap-up

   ●    Best practices and troubleshooting
   ●    Resources for continued learning