Mateusz Ochal

Research project title: 
Few-Shot Learning for Underwater Image Classification
Research project: 

Object recognition on underwater optical and sonar images faces many challenges. Finding an algorithm that can quickly and effectively learn from limited data in these demanding environments would alleviate the need for collecting and manually labelling huge datasets, potentially lowering down the cost of underwater missions, and making deep-sea exploration more available. My research investigates Few-Shot Learning (FSL) algorithms in a practical underwater setting, where deep learning models are trained with limited data.

Student type: 
Current student