Latest Updates
NLDL Conference 2025 - Tromsø, Norway
Presenting research on fine-grained fish re-identification using Vision Transformers at the Northern Lights Deep Learning conference. Discussing advances in metric learning for fisheries monitoring.
Paper Accepted: Fine-Grained Fish Classification
Our paper "Towards Visual Re-Identification of fish using Fine-Grained Classification for Electronic Monitoring in Fisheries" has been accepted for publication. Focus on species identification accuracy for commercial fisheries.
Data Collection in Dewatering systems
Successfully completed the initial data collection round for the OptiFish project where several video data were collected in different camera settings. The data collection was done using the dewatering system in the lab
Research & Development
Current Tasks & Progress
- Core goal: Developing robust Electronic monitoring systems for pelagic vessels which can accurately detect bycatch in high-speed motion of fish on dewatering systems
- Exploring camera selection for proper data collection in pelagic dewatering systems
- Collecting data using a scaled model of De-watering system in the lab
- Data annotation for proper training of deep learning models to cover different bycatch scenarios
Current Tasks & Progress
- Developing deep learning leveraged computer vision pipelines to adaptively increase the number of species classes that can be recognized.
More Updates Coming Soon
Project images and detailed progress will be added as TEFIMO advances. Stay tuned for significant developments!
Current Tasks & Progress
- Bycatch detection algorithms in gill net fisheries
More Updates Coming Soon
Project visuals and progress documentation will be shared soon. This project is in active development!