Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain

Early detection of plant diseases is crucial for agro-holdings, farmers, and smallholders. Various neural network architectures and training methods have been employed to identify optimal solutions for plant disease classification. However, research applying one-shot or few-shot learning approaches,...

Full description

Saved in:
Bibliographic Details
Main Author: Alexander Uzhinskiy
Format: Article
Language:English
Published: MDPI AG 2025-01-01
Series:Biology
Subjects:
Online Access:https://www.mdpi.com/2079-7737/14/1/99
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832589074802671616
author Alexander Uzhinskiy
author_facet Alexander Uzhinskiy
author_sort Alexander Uzhinskiy
collection DOAJ
description Early detection of plant diseases is crucial for agro-holdings, farmers, and smallholders. Various neural network architectures and training methods have been employed to identify optimal solutions for plant disease classification. However, research applying one-shot or few-shot learning approaches, based on similarity determination, to the plantdisease classification domain remains limited. This study evaluates different loss functions used in similarity learning, including Contrastive, Triplet, Quadruplet, SphereFace, CosFace, and ArcFace, alongside various backbone networks, such as MobileNet, EfficientNet, ConvNeXt, and ResNeXt. Custom datasets of real-life images, comprising over 4000 samples across 68 classes of plant diseases, pests, and their effects, were utilized. The experiments evaluate standard transfer learning approaches alongside similarity learning methods based on two classes of loss function. Results demonstrate the superiority of cosine-based methods over Siamese networks in embedding extraction for disease classification. Effective approaches for model organization and training are determined. Additionally, the impact of data normalization is tested, and the generalization ability of the models is assessed using a special dataset consisting of 400 images of difficult-to-identify plant disease cases.
format Article
id doaj-art-c53236592ae44caa9d3d16ee3f6f3dde
institution Kabale University
issn 2079-7737
language English
publishDate 2025-01-01
publisher MDPI AG
record_format Article
series Biology
spelling doaj-art-c53236592ae44caa9d3d16ee3f6f3dde2025-01-24T13:23:39ZengMDPI AGBiology2079-77372025-01-011419910.3390/biology14010099Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification DomainAlexander Uzhinskiy0Joint Institute for Nuclear Research, 6 Joliot-Curie, Dubna 141980, RussiaEarly detection of plant diseases is crucial for agro-holdings, farmers, and smallholders. Various neural network architectures and training methods have been employed to identify optimal solutions for plant disease classification. However, research applying one-shot or few-shot learning approaches, based on similarity determination, to the plantdisease classification domain remains limited. This study evaluates different loss functions used in similarity learning, including Contrastive, Triplet, Quadruplet, SphereFace, CosFace, and ArcFace, alongside various backbone networks, such as MobileNet, EfficientNet, ConvNeXt, and ResNeXt. Custom datasets of real-life images, comprising over 4000 samples across 68 classes of plant diseases, pests, and their effects, were utilized. The experiments evaluate standard transfer learning approaches alongside similarity learning methods based on two classes of loss function. Results demonstrate the superiority of cosine-based methods over Siamese networks in embedding extraction for disease classification. Effective approaches for model organization and training are determined. Additionally, the impact of data normalization is tested, and the generalization ability of the models is assessed using a special dataset consisting of 400 images of difficult-to-identify plant disease cases.https://www.mdpi.com/2079-7737/14/1/99plant disease classificationdeep learningfew-shot learningone-shot learningSiamese networksangular margin-based loss function
spellingShingle Alexander Uzhinskiy
Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
Biology
plant disease classification
deep learning
few-shot learning
one-shot learning
Siamese networks
angular margin-based loss function
title Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
title_full Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
title_fullStr Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
title_full_unstemmed Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
title_short Evaluation of Different Few-Shot Learning Methods in the Plant Disease Classification Domain
title_sort evaluation of different few shot learning methods in the plant disease classification domain
topic plant disease classification
deep learning
few-shot learning
one-shot learning
Siamese networks
angular margin-based loss function
url https://www.mdpi.com/2079-7737/14/1/99
work_keys_str_mv AT alexanderuzhinskiy evaluationofdifferentfewshotlearningmethodsintheplantdiseaseclassificationdomain