Morphometric traits from images accurately estimate body weight of intensively reared European seabass (<i>Dicentrarchus labrax</i>)

Automated body weight (BW) estimation of intensively reared fish is a main goal of precision fish farming that can foster efficient monitoring and management of farms. Computer vision-based systems show a rapid development and require models estimating fish BW from images. The present study aimed to evaluate BW estimation models for European seabass (<i>Dicentrarchus labrax</i>) of 50 to 1,000 g. To this end, 2,933 intensively reared European seabass were individually weighed and photographed to measure several morphometric traits. Traits showing strong association with BW were included in regression analyses. BW was accurately estimated by power, multiple linear and 2nd-order polynomial models (coefficient of determination <i>R</i>² &gt; 0.9820, mean absolute percentage error from 5.7 to 6.7%) with body area, standard length, and/or body height as predictors. Present results can help to develop robust software tools for computer vision-based systems and direct measurement of BW of European seabass from images.