Distinguishing Lumpy Skin Disease from Coat Patterns Using Morph-ological Priors in Deep Learning.

Morphology-driven AI screening tool aids early LSD detection and herd-level monitoring.

Lumpy skin disease (LSD) is a significant transboundary cattle disease causing substantial economic and welfare losses. Early detection on-farm is challenging due to small lesions, poor contrast, and obscuration by coat patterns. This study presents LSDD, a morphology-driven deep learning model for automated LSD detection using on-farm camera and smartphone images. The system incorporates two key modules: a local texture refinement module to identify subtle nodules while reducing background noise, and a global morphology consistency module to recognize circular or clustered lesion configurations while minimizing false positives from elongated coat markings. The model provides binary image-level classification (healthy versus lesion) and was validated across diverse farm conditions including varying illumination, backgrounds, and coat patterns. Results demonstrate LSDD's practical utility as a herd-level screening tool to identify animals with LSD-compatible lesions earlier, facilitating timely veterinary intervention and supporting enhanced on-farm surveillance and control measures. Authors emphasize this tool complements rather than replaces clinical examination and laboratory confirmation.

This summary was distilled by AI and may occasionally misinterpret data. Confirm critical details with the primary literature before clinical application.