What is iMouse? iMouse stands for "Integrated Monitoring & Operation System in Use Within a Standardised Environment". It is a state-of-the-art digitalisation solution for research institutions that can be used to observe mice in preclinical research and study their behaviour using AI.
The SIBB e.V. digitisation approach: The system uses cameras in combination with an invisible night light to monitor standard home cages and enable unbiased observation of the animals around the clock. This leads to objective results, which are crucial for universities, pharmaceutical companies and the chemical industry. By law, drugs must be tested on animals before they can be authorised for human use. The aim of SIBB e.V. is to improve animal welfare, achieve more precise results and ultimately reduce the number of laboratory animals required.
Open innovation and open source: SIBB e.V. works according to the principles of open innovation and open source and co-operates with several universities that use the product for their research. These partnerships help SIBB to continuously improve its machine learning models. SIBB e.V. aims to build a community and connect ML experts and researchers in Berlin and Germany to foster the open innovation landscape.
Hackathon agenda:
Day 1:
- Introduction to the system (hardware & software)
- Goals of the system - Why develop a Digital Home Cage Monitoring System (DHCMS)?
- Challenges during the development
- Introduction to the AI process
- Brainstorming to optimise the AI process
- Practical exercises, group formation and team building
Day 2:
- Recap of day 1 and discussion of the challenges
- Results of day 1
- Further discussion points
- Regrouping and practical exercises
- Results of the day
- Conclusion, feedback round and future communication
Further discussion points:
- Codec selection to reduce bandwidth for scaling
- Synchronisation of multiple cameras
- Use of Yolov8, Yolov10, and Autodistill
- Integration of Kafka
- Improving classification accuracy with CLIP or BLIP
- Options for cloud acceleration
- Platform for external annotation and quality control
- Collaboration with the scientific community
- Implementation of a remote maintenance platform for real-time hardware status
The goals of the hackathon:
- Develop an optimised ML process to detect behaviours such as drinking, eating, standing, climbing, sleeping and moving.
- Hiring a senior AI contributor for the iMouse team.
What's on offer:
- Free food and drinks
- Insights into the preclinical research environment
- Real video footage from labs for AI training
- Computing power sponsored by our partner
Join this exciting opportunity in Berlin to make a real impact on preclinical research through AI innovation!
SIBB - Verband der Software-, Informations- und Kommunikations-Industrie in Berlin und Brandenburg e.V.
Oranienplatz 4, 10999 Berlin