Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Simultaneous Detection of LSD and FMD in Cattle Using Ensemble Deep Learning

Nazibul Basar Ayon, Abdul Hasib, Md. Faishal Ahmed, Md. Sadiqur Rahman, Kamrul Islam, T. M. Mehrab Hasan, A. S. M. Ahsanul Sarkar Akib

TL;DR

This work tackles the urgent need for automated, accurate detection of two cattle diseases, LSD and FMD, which share overlapping symptoms. It introduces a six-class, ensemble deep-learning framework combining VGG16, ResNet50, and InceptionV3 with weighted averaging and probability calibration to overcome symptom overlap. Leveraging a large, expert-annotated dataset of 10,516 images from 18 farms across three countries and supplemented by cGAN-generated synthetic data, the approach achieves 98.2% accuracy and a 99.5% AUC-ROC, with strong per-class performance and high specificity. The results demonstrate practical potential for rapid, automated veterinary diagnostics in resource-constrained settings, while outlining future expansions to diverse geographies, edge deployment, and field validation to ensure real-world impact.

Abstract

Lumpy Skin Disease (LSD) and Foot-and-Mouth Disease (FMD) are highly contagious viral diseases affecting cattle, causing significant economic losses and welfare challenges. Their visual diagnosis is complicated by significant symptom overlap with each other and with benign conditions like insect bites or chemical burns, hindering timely control measures. Leveraging a comprehensive dataset of 10,516 expert-annotated images from 18 farms across India, Brazil, and the USA, this study presents a novel Ensemble Deep Learning framework integrating VGG16, ResNet50, and InceptionV3 with optimized weighted averaging for simultaneous LSD and FMD detection. The model achieves a state-of-the-art accuracy of 98.2\%, with macro-averaged precision of 98.2\%, recall of 98.1\%, F1-score of 98.1\%, and an AUC-ROC of 99.5\%. This approach uniquely addresses the critical challenge of symptom overlap in multi-disease detection, enabling early, precise, and automated diagnosis. This tool has the potential to enhance disease management, support global agricultural sustainability, and is designed for future deployment in resource-limited settings.

Simultaneous Detection of LSD and FMD in Cattle Using Ensemble Deep Learning

TL;DR

This work tackles the urgent need for automated, accurate detection of two cattle diseases, LSD and FMD, which share overlapping symptoms. It introduces a six-class, ensemble deep-learning framework combining VGG16, ResNet50, and InceptionV3 with weighted averaging and probability calibration to overcome symptom overlap. Leveraging a large, expert-annotated dataset of 10,516 images from 18 farms across three countries and supplemented by cGAN-generated synthetic data, the approach achieves 98.2% accuracy and a 99.5% AUC-ROC, with strong per-class performance and high specificity. The results demonstrate practical potential for rapid, automated veterinary diagnostics in resource-constrained settings, while outlining future expansions to diverse geographies, edge deployment, and field validation to ensure real-world impact.

Abstract

Lumpy Skin Disease (LSD) and Foot-and-Mouth Disease (FMD) are highly contagious viral diseases affecting cattle, causing significant economic losses and welfare challenges. Their visual diagnosis is complicated by significant symptom overlap with each other and with benign conditions like insect bites or chemical burns, hindering timely control measures. Leveraging a comprehensive dataset of 10,516 expert-annotated images from 18 farms across India, Brazil, and the USA, this study presents a novel Ensemble Deep Learning framework integrating VGG16, ResNet50, and InceptionV3 with optimized weighted averaging for simultaneous LSD and FMD detection. The model achieves a state-of-the-art accuracy of 98.2\%, with macro-averaged precision of 98.2\%, recall of 98.1\%, F1-score of 98.1\%, and an AUC-ROC of 99.5\%. This approach uniquely addresses the critical challenge of symptom overlap in multi-disease detection, enabling early, precise, and automated diagnosis. This tool has the potential to enhance disease management, support global agricultural sustainability, and is designed for future deployment in resource-limited settings.
Paper Structure (19 sections, 9 equations, 5 figures, 8 tables, 2 algorithms)

This paper contains 19 sections, 9 equations, 5 figures, 8 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Methodology
  4. Dataset Description & Data Acquisition
  5. Data Preprocessing
  6. Individual Model Architectures
  7. Individual Model Training
  8. Ensemble Model and Training
  9. Evaluation Preparation
  10. Results & Analysis
  11. Individual Model Performance
  12. Ablation Study
  13. Per-Class Performance
  14. Model Performance Evaluation
  15. Training Dynamics
  16. ...and 4 more sections

Figures (5)

  • Figure 1: Methodology flowchart.
  • Figure 2: Ensemble architecture integrating base model predictions.
  • Figure 3: Ensemble model performance metrics: (a) Confusion matrix shows only 37 misclassifications out of 2,052 test samples; (b) ROC curves demonstrate near-perfect class separation with 99.5% macro-averaged AUC-ROC
  • Figure 4: Learning curves over 200 epochs: (a) Training accuracy improves from 85% to 98.2%; (b) Validation loss decreases from 0.3 to 0.038, demonstrating stable convergence without overfitting
  • Figure 5: Training dynamics over selected epochs.