HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot

Soybean frogeye leaf spot (FLS), a serious soybean disease, causes severe yield losses in the largest production regions of China. However, both conventional field monitoring and machine learning algorithms remain challenged in achieving rapid and accurate detection. In this study, an HSDT-TabNet mo...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaoming Li, Yang Zhou, Yongguang Li, Shiqi Wang, Wenxue Bian, Hongmin Sun
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Agronomy
Subjects:
Online Access:https://www.mdpi.com/2073-4395/15/7/1530
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850067574259187712
author Xiaoming Li
Yang Zhou
Yongguang Li
Shiqi Wang
Wenxue Bian
Hongmin Sun
author_facet Xiaoming Li
Yang Zhou
Yongguang Li
Shiqi Wang
Wenxue Bian
Hongmin Sun
author_sort Xiaoming Li
collection DOAJ
description Soybean frogeye leaf spot (FLS), a serious soybean disease, causes severe yield losses in the largest production regions of China. However, both conventional field monitoring and machine learning algorithms remain challenged in achieving rapid and accurate detection. In this study, an HSDT-TabNet model was proposed for the grading of soybean FLS under field conditions by analyzing unmanned aerial vehicle (UAV)-based hyperspectral data. This model employs a dual-path parallel feature extraction strategy: the TabNet path performs sparse feature selection to capture fine-grained local discriminative information, while the hierarchical soft decision tree (HSDT) path models global nonlinear relationships across hyperspectral bands. The features from both paths are then dynamically fused via a multi-head attention mechanism to integrate complementary information. Furthermore, the overall generalization ability of the model is improved through hyperparameter optimization based on the tree-structured Parzen estimator (TPE). Experimental results show that HSDT-TabNet achieved a macro-accuracy of 96.37% under five-fold cross-validation. It outperformed the TabTransformer and SVM baselines by 2.08% and 2.23%, respectively. For high-severity cases (Level 4–5), the classification accuracy exceeded 97%. This study provides an effective method for precise field-scale crop disease monitoring.
format Article
id doaj-art-9b3a1b96bc5c432aa0c455cbef204391
institution DOAJ
issn 2073-4395
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Agronomy
spelling doaj-art-9b3a1b96bc5c432aa0c455cbef2043912025-08-20T02:48:16ZengMDPI AGAgronomy2073-43952025-06-01157153010.3390/agronomy15071530HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf SpotXiaoming Li0Yang Zhou1Yongguang Li2Shiqi Wang3Wenxue Bian4Hongmin Sun5College of Electrical Engineering and Information, Northeast Agricultural University, Harbin 150030, ChinaCollege of Electrical Engineering and Information, Northeast Agricultural University, Harbin 150030, ChinaCollege of Agriculture, Northeast Agricultural University, Harbin 150030, ChinaCollege of Electrical Engineering and Information, Northeast Agricultural University, Harbin 150030, ChinaCollege of Electrical Engineering and Information, Northeast Agricultural University, Harbin 150030, ChinaCollege of Electrical Engineering and Information, Northeast Agricultural University, Harbin 150030, ChinaSoybean frogeye leaf spot (FLS), a serious soybean disease, causes severe yield losses in the largest production regions of China. However, both conventional field monitoring and machine learning algorithms remain challenged in achieving rapid and accurate detection. In this study, an HSDT-TabNet model was proposed for the grading of soybean FLS under field conditions by analyzing unmanned aerial vehicle (UAV)-based hyperspectral data. This model employs a dual-path parallel feature extraction strategy: the TabNet path performs sparse feature selection to capture fine-grained local discriminative information, while the hierarchical soft decision tree (HSDT) path models global nonlinear relationships across hyperspectral bands. The features from both paths are then dynamically fused via a multi-head attention mechanism to integrate complementary information. Furthermore, the overall generalization ability of the model is improved through hyperparameter optimization based on the tree-structured Parzen estimator (TPE). Experimental results show that HSDT-TabNet achieved a macro-accuracy of 96.37% under five-fold cross-validation. It outperformed the TabTransformer and SVM baselines by 2.08% and 2.23%, respectively. For high-severity cases (Level 4–5), the classification accuracy exceeded 97%. This study provides an effective method for precise field-scale crop disease monitoring.https://www.mdpi.com/2073-4395/15/7/1530soybeanfrogeye leaf spotHSDT-TabNethyperspectral reflectanceUAV
spellingShingle Xiaoming Li
Yang Zhou
Yongguang Li
Shiqi Wang
Wenxue Bian
Hongmin Sun
HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
Agronomy
soybean
frogeye leaf spot
HSDT-TabNet
hyperspectral reflectance
UAV
title HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
title_full HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
title_fullStr HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
title_full_unstemmed HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
title_short HSDT-TabNet: A Dual-Path Deep Learning Model for Severity Grading of Soybean Frogeye Leaf Spot
title_sort hsdt tabnet a dual path deep learning model for severity grading of soybean frogeye leaf spot
topic soybean
frogeye leaf spot
HSDT-TabNet
hyperspectral reflectance
UAV
url https://www.mdpi.com/2073-4395/15/7/1530
work_keys_str_mv AT xiaomingli hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot
AT yangzhou hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot
AT yongguangli hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot
AT shiqiwang hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot
AT wenxuebian hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot
AT hongminsun hsdttabnetadualpathdeeplearningmodelforseveritygradingofsoybeanfrogeyeleafspot