Automated assessment of cardiac morphological variation in Atlantic salmon (Salmo salar L.)

Deviating heart shapes and poor cardiac health is a recurring concern in farmed Atlantic salmon. Morphometric analysis has so far improved our understanding of salmonid cardiac morphology, but assessment of morphological cardiac variation is usually performed manually through measurements of lengths, ratios, and angles. Manual assessment of heart shape is tedious, time-consuming, and not very standardized. It also requires training and alignment of personnel to achieve reliable results. Considering these challenges, we aimed to automate this process using a deep learning model for computer vision to measure the morphological variations of the heart. Here we developed an algorithm for a diagnostic tool to detect variation in cardiac morphology in farmed Atlantic salmon, which we believe can assess cardiac morphological variation in a more objective, reproducible, and reliable manner compared to the manual process. The knowledge derived from this study may represent a crucial step in comprehending and eventually reducing cardiac abnormalities in farmed salmonids, which is essential for improving fish health and welfare and ensuring aquaculture's sustainable growth.