Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification
Deep learning with convolutional neural networks represents the most used approach in recent years in the classification of leaves’ diseases. The literature has extensively addressed the problem using laboratory-acquired datasets with a homogeneous background. In this article, we explore the variabi...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-09-01
|
Series: | Agronomy |
Subjects: | |
Online Access: | https://www.mdpi.com/2073-4395/12/10/2359 |
_version_ | 1797476151477665792 |
---|---|
author | Gianni Fenu Francesca Maridina Malloci |
author_facet | Gianni Fenu Francesca Maridina Malloci |
author_sort | Gianni Fenu |
collection | DOAJ |
description | Deep learning with convolutional neural networks represents the most used approach in recent years in the classification of leaves’ diseases. The literature has extensively addressed the problem using laboratory-acquired datasets with a homogeneous background. In this article, we explore the variability factors that influence the classification of plant diseases by analyzing the same plant and disease under different conditions, i.e., in the field and in the laboratory. Two plant species and five biotic stresses are analyzed using different architectures, such as EfficientB0, MobileNetV2, InceptionV2, ResNet50 and VGG16. Experiments show that model performance drops drastically when using representative datasets, and the features learned from the network to determine the class do not always belong to the leaf lesion. In the worst case, the accuracy drops from 92.67% to 54.41%. Our results indicate that while deep learning is an effective technique, there are some technical issues to consider when applying it to more representative datasets collected in the field. |
first_indexed | 2024-03-09T20:54:52Z |
format | Article |
id | doaj.art-ce05adf923584281a2d0c6d3d7657246 |
institution | Directory Open Access Journal |
issn | 2073-4395 |
language | English |
last_indexed | 2024-03-09T20:54:52Z |
publishDate | 2022-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Agronomy |
spelling | doaj.art-ce05adf923584281a2d0c6d3d76572462023-11-23T22:25:46ZengMDPI AGAgronomy2073-43952022-09-011210235910.3390/agronomy12102359Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease ClassificationGianni Fenu0Francesca Maridina Malloci1Department of Mathematics and Computer Science, University of Cagliari, Via Ospedale 72, 09124 Cagliari, ItalyDepartment of Mathematics and Computer Science, University of Cagliari, Via Ospedale 72, 09124 Cagliari, ItalyDeep learning with convolutional neural networks represents the most used approach in recent years in the classification of leaves’ diseases. The literature has extensively addressed the problem using laboratory-acquired datasets with a homogeneous background. In this article, we explore the variability factors that influence the classification of plant diseases by analyzing the same plant and disease under different conditions, i.e., in the field and in the laboratory. Two plant species and five biotic stresses are analyzed using different architectures, such as EfficientB0, MobileNetV2, InceptionV2, ResNet50 and VGG16. Experiments show that model performance drops drastically when using representative datasets, and the features learned from the network to determine the class do not always belong to the leaf lesion. In the worst case, the accuracy drops from 92.67% to 54.41%. Our results indicate that while deep learning is an effective technique, there are some technical issues to consider when applying it to more representative datasets collected in the field.https://www.mdpi.com/2073-4395/12/10/2359deep learningconvolutional neural networkbenchmarkplant disease prediction |
spellingShingle | Gianni Fenu Francesca Maridina Malloci Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification Agronomy deep learning convolutional neural network benchmark plant disease prediction |
title | Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification |
title_full | Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification |
title_fullStr | Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification |
title_full_unstemmed | Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification |
title_short | Evaluating Impacts between Laboratory and Field-Collected Datasets for Plant Disease Classification |
title_sort | evaluating impacts between laboratory and field collected datasets for plant disease classification |
topic | deep learning convolutional neural network benchmark plant disease prediction |
url | https://www.mdpi.com/2073-4395/12/10/2359 |
work_keys_str_mv | AT giannifenu evaluatingimpactsbetweenlaboratoryandfieldcollecteddatasetsforplantdiseaseclassification AT francescamaridinamalloci evaluatingimpactsbetweenlaboratoryandfieldcollecteddatasetsforplantdiseaseclassification |