Who I Am and What I Do

My name is Samitha N. Thilakarathna. I'm a PhD student at DTU Aqua, and I am specializing in cutting-edge artificial intelligence solutions for automated fisheries catch monitoring.

My research spans the breadth of pelagic, demersal trawl, and gillnet fisheries, targeting the development of advanced systems for species identification, quantification, and comprehensive documentation.

By harnessing AI technologies, my work strives for 100% coverage in monitoring and data accuracy—driving the transformation toward transparent, sustainable fisheries management for the future.

Watch my latest Documentary
Animated GIF of a fish sorting belt.
Annotations using Supervisely.
Camera view of fish on a conveyor.

My Background

I began my academic journey in Fisheries and Marine Science (BSc.) at the University of Ruhuna in Sri Lanka, gaining a comprehensive understanding of the fishing industry underpinned by rigorous training in mathematics and statistics.

Building on this strong foundation, I pursued a Master's degree at DTU Aqua, Technical University of Denmark, focusing on Sustainable Fisheries and Aquaculture. It was here that my fascination with artificial intelligence and computer vision first took root, culminating in a master's thesis on Fish Re-Identification using deep learning techniques—specifically, Vision Transformers.

Now, as a PhD student at DTU Aqua, my work centers on advanced computer vision and AI applications for automated fisheries monitoring. This scientific journey has been both transformative and deeply rewarding, thanks in large part to the inspiring mentorship and collaborative spirit within DTU. The future holds many exciting developments, and I am eager to contribute innovative solutions to the challenges of fisheries science.

My Learning Curve

Portrait of Samitha N. Thilakarathna
A graph showing my learning curve in AI over time.

My Articles

Fish Re-Identification

Importance of Fish Re-Identification in Fisheries Catch Documentation
Fish Species Classifier

How to train a Fish Species Classifier with Deep Learning
Fisheries Data

Types of Fisheries Data
SVO to AVI converter

SVO to AVI video converter app (For ZED camera users)
AI in Sustainable Fisheries

Why are AI applications necessary in Sustainable Fisheries?
Transfer Learning

Transfer Learning
Contrastive Learning

Contrastive Learning
Cat Experiment for Neural Networks

Cat Experiment for Modern Neural Networks

Special Projects

Offshore Wind Farms: Nurturing Fish Spawning Grounds

Explore how offshore wind farms can enhance marine biodiversity by creating ideal conditions for fish spawning. These installations often feature monopile foundations with scour protection made of large rocks, forming reef-like structures that stabilize the seabed and prevent erosion. These artificial reefs provide habitats that boost fish abundance, growth, and survival. Additionally, fishing restrictions within wind farm areas can transform them into sanctuaries for various marine species. This video delves into a study by DTU Aqua investigating whether Atlantic cod (Gadus morhua) spawn within a Danish offshore wind farm in the North Sea. Discover how such environments may support marine life and contribute to the restoration of sustainable fish populations across European waters.

View Publication in DTU Orbit

My Skills

Over the past years I have gained and practiced some extra skills that became useful in my academic journey.

Video Production
Digital Illustration
2D/3D Animation
Programming (R, Python)
Adobe Photoshop Adobe After Effects Adobe Premiere Pro Adobe Illustrator Blender Python R Programming
Image of a project
Image of another project
Image of a fourth project
Image of a Fifth project
Image of a third project