A Deep Learning Approach for Detection and Localization of Leaf Anomalies

Davide Calabrò; Massimiliano Lupo Pasini; Nicola Ferro; Simona Perotto

doi:10.1007/978-3-031-55060-7_3

A Deep Learning Approach for Detection and Localization of Leaf Anomalies

Davide Calabrò, Massimiliano Lupo Pasini, Nicola Ferro, Simona Perotto

Computational Coupled Physics

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

Abstract

The detection and localization of possible diseases in crops are usually automated by resorting to supervised deep learning approaches. In this work, we tackle these goals with unsupervised models, by applying three different types of autoencoders to a specific open-source dataset of healthy and unhealthy pepper and cherry leaf images. CAE, CVAE and VQ-VAE autoencoders are deployed to screen unlabeled images of such a dataset, and compared in terms of image reconstruction, anomaly removal, detection and localization. The vector-quantized variational architecture turns out to be the best performing one with respect to all these targets.

Original language	English
Title of host publication	Lecture Notes in Computational Science and Engineering
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	43-66
Number of pages	24
DOIs	https://doi.org/10.1007/978-3-031-55060-7_3
State	Published - 2024

Publication series

Name	Lecture Notes in Computational Science and Engineering
Volume	151
ISSN (Print)	1439-7358
ISSN (Electronic)	2197-7100

Funding

Massimiliano Lupo Pasini thanks Dr. Vladimir Protopopescu for his valuable feedback in the preparation of this manuscript. Massimiliano Lupo Pasini\u2019s work was supported in part by the Artificial Intelligence Initiative as part of the Laboratory Directed Research and Development (LDRD) Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725. Simona Perotto and Nicola Ferro acknowledge the support by MUR, grant Dipartimento di Eccellenza 2023\u20132027. Acknowledgements Massimiliano Lupo Pasini thanks Dr. Vladimir Protopopescu for his valuable feedback in the preparation of this manuscript. Massimiliano Lupo Pasini\u2019s work was supported in part by the Artificial Intelligence Initiative as part of the Laboratory Directed Research and Development (LDRD) Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725. Simona Perotto and Nicola Ferro acknowledge the support by MUR, grant Dipartimento di Eccellenza 2023\u20132027.

Access to Document

10.1007/978-3-031-55060-7_3

Cite this

Calabrò, D., Pasini, M. L., Ferro, N., & Perotto, S. (2024). A Deep Learning Approach for Detection and Localization of Leaf Anomalies. In Lecture Notes in Computational Science and Engineering (pp. 43-66). (Lecture Notes in Computational Science and Engineering; Vol. 151). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-55060-7_3

@inbook{ffcd27c85b194607a0e8ba164a21593e,

title = "A Deep Learning Approach for Detection and Localization of Leaf Anomalies",

abstract = "The detection and localization of possible diseases in crops are usually automated by resorting to supervised deep learning approaches. In this work, we tackle these goals with unsupervised models, by applying three different types of autoencoders to a specific open-source dataset of healthy and unhealthy pepper and cherry leaf images. CAE, CVAE and VQ-VAE autoencoders are deployed to screen unlabeled images of such a dataset, and compared in terms of image reconstruction, anomaly removal, detection and localization. The vector-quantized variational architecture turns out to be the best performing one with respect to all these targets.",

author = "Davide Calabr{\`o} and Pasini, {Massimiliano Lupo} and Nicola Ferro and Simona Perotto",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.",

year = "2024",

doi = "10.1007/978-3-031-55060-7_3",

language = "English",

series = "Lecture Notes in Computational Science and Engineering",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "43--66",

booktitle = "Lecture Notes in Computational Science and Engineering",

}

A Deep Learning Approach for Detection and Localization of Leaf Anomalies. / Calabrò, Davide; Pasini, Massimiliano Lupo; Ferro, Nicola et al.
Lecture Notes in Computational Science and Engineering. Springer Science and Business Media Deutschland GmbH, 2024. p. 43-66 (Lecture Notes in Computational Science and Engineering; Vol. 151).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

TY - CHAP

T1 - A Deep Learning Approach for Detection and Localization of Leaf Anomalies

AU - Calabrò, Davide

AU - Pasini, Massimiliano Lupo

AU - Ferro, Nicola

AU - Perotto, Simona

PY - 2024

Y1 - 2024

N2 - The detection and localization of possible diseases in crops are usually automated by resorting to supervised deep learning approaches. In this work, we tackle these goals with unsupervised models, by applying three different types of autoencoders to a specific open-source dataset of healthy and unhealthy pepper and cherry leaf images. CAE, CVAE and VQ-VAE autoencoders are deployed to screen unlabeled images of such a dataset, and compared in terms of image reconstruction, anomaly removal, detection and localization. The vector-quantized variational architecture turns out to be the best performing one with respect to all these targets.

AB - The detection and localization of possible diseases in crops are usually automated by resorting to supervised deep learning approaches. In this work, we tackle these goals with unsupervised models, by applying three different types of autoencoders to a specific open-source dataset of healthy and unhealthy pepper and cherry leaf images. CAE, CVAE and VQ-VAE autoencoders are deployed to screen unlabeled images of such a dataset, and compared in terms of image reconstruction, anomaly removal, detection and localization. The vector-quantized variational architecture turns out to be the best performing one with respect to all these targets.

UR - http://www.scopus.com/inward/record.url?scp=85200471450&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-55060-7_3

DO - 10.1007/978-3-031-55060-7_3

M3 - Chapter

AN - SCOPUS:85200471450

T3 - Lecture Notes in Computational Science and Engineering

SP - 43

EP - 66

BT - Lecture Notes in Computational Science and Engineering

PB - Springer Science and Business Media Deutschland GmbH

ER -