A New Domain-Independent Approach for Classification of Bacteria, Fungus, and Virus-Infected Fruit and Leaf Images

Abstract

Early and reliable detection of bacterial, fungal, and viral infections in fruits and leaves is essential for improving crop productivity, preventing disease spread, and supporting food security. Most existing approaches are domain-specific and struggle to generalize across diverse plant organs or varying image qualities. To address this challenge, we propose a novel domain-independent classification framework that integrates quality-metric features—Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM)—with an adapted lightweight Convolutional Neural Network (CNN). This is the first approach that explores quality measures as features for addressing challenges of classification of fruits and leaves infected by virus, fungus, and bacteria. The method first performs connected-component analysis on K-means clusters generated from R, G, B, and Gray channels to isolate disease-relevant regions and extract quality-based features. These features are fused with visual features extracted from the RGB images using a multimodal CNN architecture. Extensive experiments conducted on the proposed fruit–leaf dataset and four external benchmark datasets demonstrate that the model achieves high accuracy, strong robustness to blur, noise, rotation, and scaling, and superior generalization performance compared with state-of-the-art methods. Cross-domain evaluations further confirm that the proposed method is domain-independent and reliable for the classification of fruits and leaves infected by bacteria, fungi, and viruses.

Received: 24 September 2025 | Revised: 2 December 2025 | Accepted: 25 December 2025

Conflicts of Interest

Palaiahnakote Shivakumara is the Editor-in-Chief for Artificial Intelligence and Applications, and he is not involved in the editorial review or the decision to publish this article. The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Author Contribution Statement

Poornima Basatti Hanuma Gowda: Conceptualization, Software, Data curation, Writing – original draft, Visualization. Basavanna Mahadevappa: Formal analysis, Investigation, Supervision, Project administration. Shivakumara Palaiahnakote: Methodology, Writing – review & editing. Muhammad Hammad Saleem: Validation. Niranjan Mallappa Hanumanthu: Resources.