🎓 Xabier Lekunberri defends his PhD on Computer Vision for Sustainable Fisheries Management!

🎉 Congratulations to Xabier Lekunberri on the successful defense of his PhD thesis,
“Computer Vision for Sustainable Fisheries Management”, on Thursday, February 27, 2026 (10:00 am) at the Research Centre for Experimental Marine Biology & Biotechnology (PiE, EHU)!

His thesis was supervised by Dr. Ignacio Arganda-Carreras (University of the Basque Country, UPV/EHU) and Dr. Jose A. Fernandes-Salvador (AZTI).

🧠 Thesis Contributions

Xabier’s PhD work focuses on developing deep learning and computer vision models for monitoring and managing marine resources, addressing critical challenges in sustainable fisheries and aquaculture management. His research tackles the regulatory compliance demands in the European Union and demonstrates how AI can support transparency, efficiency, and ecological responsibility in marine resource exploitation.

Key contributions include:

• Automated Tuna Catch Composition Estimation
  An innovative multi-stage pipeline using Mask R-CNN for segmentation and fine-tuned ResNet50 for classification of tuna species in electronic monitoring (EM) camera footage from purse seine vessels. The final pipeline segments 84.8% of individuals in test videos with a mean average error of 4.5%.

• Video Processing and Individual Tracking
  Extended the pipeline to handle video data rather than static images, enabling real-time tracking of individuals within the camera’s field of view. A new ground truth dataset created through extensive sampling by scientists and observers significantly improved model accuracy.

• SAR-Based Aquaculture Farm Detection
  Developed a pipeline using Sentinel-1 SAR satellite data combined with the iDPolRAD algorithm and deep learning classifiers to automatically detect and quantify aquaculture farm deployments in the North East Atlantic Ocean. Achieved 93% accuracy using a Vision Transformer-informed approach.

His work demonstrates how scalable computer vision technologies can deliver near real-time monitoring capabilities that support regulatory compliance, scientific research, and industry best practices for sustainable marine resource management.

🧑‍⚖️ Defense Committee

The committee for his defense was composed of:

• Aitor Almeida Escondrillas (Chair - Lehendakaria, DeutoTech, University of Deusto, Spain)
• Naiara Aginako Bengoa (Secretary - Idazkaria, UPV/EHU, Spain)
• Gala Moreno Arriola (Voting member - Kidea, ISSF)

👏 Bravo, Xabier! Your research makes significant contributions to sustainable marine resource management through cutting-edge computer vision and deep learning. Your work demonstrates how advanced AI technologies can support fisheries monitoring, compliance, and ecological responsibility. We wish you the best for your next steps! 🚀