Tiny deep learning model for insect segmentation and counting on resource-constrained devices

Simple item page
dc.contributor.authorKargar, Aminen
dc.contributor.authorZorbas, Dimitriosen
dc.contributor.authorGaffney, Michaelen
dc.contributor.authorO'Flynn, Brendanen
dc.contributor.authorTedesco, Salvatoreen
dc.contributor.funderDepartment of Agriculture, Food and the Marineen
dc.contributor.funderIrish Food and Agriculture Authorityen
dc.contributor.funderEuropean Regional Development Funden
dc.date.accessioned2025-04-30T08:44:42Z
dc.date.available2025-04-30T08:44:42Z
dc.date.issued2025en
dc.description.abstractAutomated insect monitoring is essential for early detection of insect pest infestations in orchards. It helps growers make an informed decision to control the insect pest population in their fields to avoid crop losses and improve crop quality. This study proposed a tiny deep-learning model for insect segmentation and counting suitable for battery-powered microcontroller (MCU)-based edge devices. For this aim, the critical layers in terms of peak memory usage in a U-Net-inspired model were recognized and then optimized to meet resources constraints on an MCU with 1 MB of RAM and 2 MB of flash storage. We then introduced an image dataset for the insect of interest, Halyomorpha halys, and the dataset-splitting strategy for the model training. The proposed model was investigated from different aspects to evaluate its performance and the feasibility of its implementation on battery-powered MCU-based edge devices. The proposed deep learning model only needs approximately 900 KB of RAM and 964 KB of storage to perform its computations and store its parameters, respectively, making it useful on edge, resource constrained systems. Moreover, each inference on an MCU-based board requires 2.6 s and consumes 4.9 J. In terms of segmentation, it achieved a Dice Similarity Coefficient (DSC) of 85% and an Intersection over Union (IoU) of 73% with a precision and recall of 83% and 86%, respectively. In terms of counting, it achieved a Mean Square Error (MSE) of 1.32, Mean Absolute Error (MAE) of 0.78 and R2 of 0.97.en
dc.description.sponsorshipEuropean Regional Development Fund (2020EN508)en
dc.description.statusPeer revieweden
dc.description.versionPublished Versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.articleid110378en
dc.identifier.citationKargar, A., Zorbas, D., Gaffney, M., O’Flynn, B. and Tedesco, S. (2025) 'Tiny deep learning model for insect segmentation and counting on resource-constrained devices', Computers and Electronics in Agriculture, 236, p.110378. DOI: 10.1016/j.compag.2025.110378en
dc.identifier.doi10.1016/j.compag.2025.110378en
dc.identifier.issn1681699en
dc.identifier.journaltitleComputers and Electronics in Agricultureen
dc.identifier.urihttps://hdl.handle.net/10468/17362
dc.identifier.volume236
dc.language.isoenen
dc.publisherElsevier B.V.en
dc.relation.project13/RC/2077-CONNECT
dc.relation.project12/RC/ 2289-P2-INSIGHT2
dc.relation.project21/RC/10303_P2-VISTAMILK2
dc.relation.project21/RC/10303_P2-VISTAMILK2which
dc.relation.project2020EN508
dc.rights© 2025, the Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)en
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectEdge computingen
dc.subjectHalyomorpha halysen
dc.subjectImage segmentationen
dc.subjectInsect monitoringen
dc.subjectMicrocontrollersen
dc.subjectOn-device deep learningen
dc.subjectTiny deep learningen
dc.titleTiny deep learning model for insect segmentation and counting on resource-constrained devicesen
dc.typeArticle (peer-reviewed)en
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
1-s2.0-S0168169925004843-main.pdf
Size:
1.81 MB
Format:
Adobe Portable Document Format
Description:
Published Version
Download
Collections
Tyndall National Institute - Journal Articles