Lumpy Skin Disease Detection in Cattle by a Robust Approach using Advanced Convolutional Neural Networks

Abstract

Background: Lumpy skin disease (LSD) is a significant health concern for cattle globally and poses economic threats by affecting various aspects of cattle health. Integrating artificial intelligence (AI) and machine learning (ML) with visual inspections and biosensor data has shown promise in enhancing disease detection and diagnosis. The present study harnesses the potential of Convolutional Neural Networks (CNN) and image processing for detecting LSD. Methods: Using images from the agricultural landscape, this study highlights the significance of convolutional neural networks that identify the lumpy skin disease (LSD) in animals. Images are categorized into two groups: LSD (infected skin) and non-LSD (normal skin). This is achieved by applying a deeply designed deep learning model carefully built to fulfill this particular need. Evaluation metrics assess the model’s performance, including accuracy, loss and a confusion matrix. Result: A CNN-based model trained for 50 epochs to classify skin conditions, achieved an 86.54% accuracy. The study underscores the potential of CNN in early LSD detection, paving the way for practical applications in veterinary medicine. Future work involves addressing dataset limitations, refining model parameters, reducing image noise, exploring different feature extraction methods and investigating additional animal skin conditions.