Reply to Mahler, S. Comment on “Jia et al. Exploratory Study of Sex Identification for Chicken Embryos Based on Blood Vessel Images and Deep Learning. Agriculture 2023, 13, 1480”

We respond to the Comment regarding the reproducibility and generalizability of in ovo sex identification from extraembryonic vessel images. Our 2023 exploratory study evaluated Jingfen No. 6 white-shell eggs using green-light transmission images and a YOLO-based object detector. Mahler et al. evaluated LSCI images from a different egg population using whole-image classifiers and reported per-egg validation accuracies below 60%. To assess what could still be verified, we reconstructed the detection workflow from retained transmission images using a new implementation based on the published methods. The reconstruction yielded validation-set mAP@0.5 values of 84.90%, 91.74%, and 80.56% for days 3, 4, and 5, respectively. These are internal validation estimates because the same partitions were used for checkpoint selection and performance reporting. Substantial differences between the studies in imaging modality, egg population, prediction task, reference labeling, and evaluation procedures preclude direct comparison of their results. The reconstructed results are specific to the retained transmission-image dataset and do not establish external validity or commercial readiness.