Automated Bioimage Analysis of Fish Morphometry: Enhancing Biomedical and Aquaculture Research Through Deep Learning

Biomedical research on skeletal disorders increasingly relies on small fish models like
zebrafish and medaka to investigate conditions such as osteoporosis and fibrous dys-
plasia. These models provide insights into human skeletal pathologies and broader
disorders like cancer and arthritis. Meanwhile, in aquaculture, farmed fish frequently
develop skeletal deformities in the jaw, operculum, and vertebral column, compromising
fish welfare, performance, and product quality. These anomalies result in significant
economic losses due to manual culling. Understanding the underlying mechanisms of
skeletal development in fish is crucial for improving both human health and aquaculture
sustainability. This thesis explores the application of deep learning methodologies in
bioimage analysis, focusing on morphometric and phenotypic studies in biomedical and
aquaculture research within the framework of EU funded BioMedAqu project