AOP: An Anti-overfitting Pretreatment for Practical Image-based Plant Diagnosis

TL;DR

This paper introduces AOP, a preprocessing method that detects plant areas and calibrates brightness to reduce overfitting in image-based plant diagnosis, significantly improving accuracy on unseen farm data.

Contribution

The study proposes AOP, a novel anti-overfitting pretreatment that enhances the robustness of plant diagnosis systems across different datasets.

Findings

AOP improves diagnosis accuracy by 12.2% on unseen farm images.

Experiments with over 50,000 cucumber leaf images validate AOP's effectiveness.

AOP reduces dataset bias and overfitting in practical plant diagnosis applications.

Abstract

In image-based plant diagnosis, clues related to diagnosis are often unclear, and the other factors such as image backgrounds often have a significant impact on the final decision. As a result, overfitting due to latent similarities in the dataset often occurs, and the diagnostic performance on real unseen data (e,g. images from other farms) is usually dropped significantly. However, this problem has not been sufficiently explored, since many systems have shown excellent diagnostic performance due to the bias caused by the similarities in the dataset. In this study, we investigate this problem with experiments using more than 50,000 images of cucumber leaves, and propose an anti-overfitting pretreatment (AOP) for realizing practical image-based plant diagnosis systems. The AOP detects the area of interest (leaf, fruit etc.) and performs brightness calibration as a preprocessing step. The experimental results demonstrate that our AOP can improve the accuracy of diagnosis for unknown test images from different farms by 12.2% in a practical setting.