Thomas Lejeune a., Fernando de las Heras b., Tiago Ferreira b., Maxence Perret b., Louis Rostaing b., Ana Hernando b., David Du Pasquier a.
a. Laboratoire Watchfrog, 1 Rue Pierre Fontaine, 91000 Evry-Courcouronnes, France
b. Bionomous SA, Z.I. du Vivier, 22 – 1690 Villaz-St-Pierre, Switzerland
ABSTRACT
Manual sorting of Medaka (Oryzias latipes) embryos for toxicology is labor-intensive and operator dependent. Following the OECD 234, OECD 240, OECD 251 and OECD 252 guidelines to medaka embryo testing, for endocrine disruption assays, requires strict validation of embryo staging and quality. Conventional manual methods for counting and sorting are not only laborious and time consuming but also provide ergonomic hazards.
The everyday counting and sorting of substantial quantities can result in musculoskeletal strain, causing discomfort in the neck, shoulders, and back among workers.
This study adapts a zebrafish-focused robotic platform for Medaka, by retraining AI models for Medaka morphology and optimizing physical handling parameters, aiming to reduce the load on personnel and minimize the time spent on counting, allowing it to be better allocated to other tasks.
"*" indicates required fields