Sexual differences in yolk images of chicken eggs captured during early hatching stages: Possibility of application for early sex prediction

• Chicken eggs in early hatching stage were studied for sexing. • Features related to growth in the embryo and vascular network were extracted. • Frequentist and Bayesian statistical inference were used for their analysis. • Sexual difference appeared in the vascular network at third day of the incubation. • The embryo area and vascular fractal dimension may be useful for sexing. Sex prediction in layer chicks before hatching is a critical issue that needs to be resolved, and vascular imaging is known to be useful. However, features distinguishing male and female chicks in the vascular image during early hatching stages remain unknown. In this study, we aimed to extract data on features related to growth in the embryo and the vascular network from yolk images captured after removing the eggshell at the third (Day 3) and fourth (Day 4) days of the incubation period. The features were analyzed using frequentist statistical inference and Bayesian inference to elucidate the appearance of sexual differences. Welch’s t -test showed some sexual differences in the embryo-vascular area and fractal dimension in Day 3 eggs ( p < 0.05). In the Bayesian inference, the fractal dimension made the greatest contribution to sex prediction, followed by the embryo-vascular area. However, these sexual differences were reduced by Day 4. Our results suggest that these features can make a sex prediction model simpler and more robust than using all transmission images.