Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Cocoa Pod Disease Classification via Transfer Learning and Ensemble Methods: Toward Robust Predictive Modeling

Devina Anduyan, Nyza Cabillo, Navy Gultiano, Mark Phil Pacot

TL;DR

This work tackles the challenge of robust cocoa pod disease classification by combining transfer learning with ensemble methods. It fine-tunes six CNN backbones (VGG16/19, ResNet50/101, InceptionV3, Xception) and evaluates Bagging, Boosting, and Stacking ensembles to detect Black Pod Rot, Pod Borer, and Healthy pods on a balanced, augmented dataset. Bagging delivers the strongest performance with perfect test accuracy on the held-out set, followed by Boosting and Stacking, indicating that ensemble diversity with transfer-learned features enhances generalization. The study demonstrates a promising path for precision agriculture and automated crop-disease management, with potential extensions to real-time monitoring and feature augmentation in field conditions.

Abstract

This study presents an ensemble-based approach for cocoa pod disease classification by integrating transfer learning with three ensemble learning strategies: Bagging, Boosting, and Stacking. Pre-trained convolutional neural networks, including VGG16, VGG19, ResNet50, ResNet101, InceptionV3, and Xception, were fine-tuned and employed as base learners to detect three disease categories: Black Pod Rot, Pod Borer, and Healthy. A balanced dataset of 6,000 cocoa pod images was curated and augmented to ensure robustness against variations in lighting, orientation, and disease severity. The performance of each ensemble method was evaluated using accuracy, precision, recall, and F1-score. Experimental results show that Bagging consistently achieved superior classification performance with a test accuracy of 100%, outperforming Boosting (97%) and Stacking (92%). The findings confirm that combining transfer learning with ensemble techniques improves model generalization and reliability, making it a promising direction for precision agriculture and automated crop disease management.

Enhancing Cocoa Pod Disease Classification via Transfer Learning and Ensemble Methods: Toward Robust Predictive Modeling

TL;DR

This work tackles the challenge of robust cocoa pod disease classification by combining transfer learning with ensemble methods. It fine-tunes six CNN backbones (VGG16/19, ResNet50/101, InceptionV3, Xception) and evaluates Bagging, Boosting, and Stacking ensembles to detect Black Pod Rot, Pod Borer, and Healthy pods on a balanced, augmented dataset. Bagging delivers the strongest performance with perfect test accuracy on the held-out set, followed by Boosting and Stacking, indicating that ensemble diversity with transfer-learned features enhances generalization. The study demonstrates a promising path for precision agriculture and automated crop-disease management, with potential extensions to real-time monitoring and feature augmentation in field conditions.

Abstract

This study presents an ensemble-based approach for cocoa pod disease classification by integrating transfer learning with three ensemble learning strategies: Bagging, Boosting, and Stacking. Pre-trained convolutional neural networks, including VGG16, VGG19, ResNet50, ResNet101, InceptionV3, and Xception, were fine-tuned and employed as base learners to detect three disease categories: Black Pod Rot, Pod Borer, and Healthy. A balanced dataset of 6,000 cocoa pod images was curated and augmented to ensure robustness against variations in lighting, orientation, and disease severity. The performance of each ensemble method was evaluated using accuracy, precision, recall, and F1-score. Experimental results show that Bagging consistently achieved superior classification performance with a test accuracy of 100%, outperforming Boosting (97%) and Stacking (92%). The findings confirm that combining transfer learning with ensemble techniques improves model generalization and reliability, making it a promising direction for precision agriculture and automated crop disease management.

Paper Structure

This paper contains 14 sections, 3 equations, 9 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Materials and methods
  4. Dataset
  5. Base Models
  6. Ensemble Learning Techniques
  7. Bagging
  8. Boosting
  9. Stacking
  10. Evaluation Metrics
  11. Experiments
  12. Cocoa pod disease classification results
  13. Model performance across ensemble methods
  14. Conclusion

Figures (9)

  • Figure 1: The conceptual framework of the proposed solution.
  • Figure 2: Sample cocoa pod images affected by Pod Borer disease.
  • Figure 3: Sample of healthy cocoa pod images.
  • Figure 4: Sample cocoa pod images affected by Black Pod Rot disease.
  • Figure 5: Confusion Matrix - Bagging (Majority Voting) Ensemble.
  • ...and 4 more figures