Publication:

Peeling Back the Pixels: Stabilizing Latent Diffusion for Data-Efficient Generation of Pineapple Fields

Date

 
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.department#PLACEHOLDER_PARENT_METADATA_VALUE#
cris.virtual.orcid0000-0002-5145-872X
cris.virtual.orcid0000-0002-9524-4205
cris.virtualsource.departmenta6e15b57-cd91-48e5-9630-242b1b7129de
cris.virtualsource.department7bac28ac-f3c2-462d-aea4-cc71c4892295
cris.virtualsource.orcida6e15b57-cd91-48e5-9630-242b1b7129de
cris.virtualsource.orcid7bac28ac-f3c2-462d-aea4-cc71c4892295
dc.contributor.authorGonzalez Hernandez, Manfred
dc.contributor.authorXie-Li, Danny
dc.contributor.authorMoya, Fabian Fallas
dc.contributor.authorRojas Gonzalez, Sebastian
dc.contributor.authorCouckuyt, Ivo
dc.date.accessioned2026-11-16T13:49:46Z
dc.date.available2026-11-16T13:49:46Z
dc.date.createdwos2026
dc.date.issued2026
dc.description.abstractWhile Latent Diffusion Models (LDMs) excel at synthetic data generation, they suffer from severe training instability on the small, homogeneous UAV datasets typical of precision agriculture. In this work, we investigate these architectural bottlenecks by training a Denoising Diffusion Probabilistic Model (DDPM) from scratch on constrained pineapple plantation imagery. Our systematic comparison reveals that continuous latent spaces (β-VAE) preserve complex crop structures, whereas discrete architectures (VQ-VAE) fail under scarce data conditions. Additionally, we demonstrate that replacing destabilizing global attention mechanisms with a purely convolutional U-Net backbone ensures smooth convergence in repetitive visual domains. Ultimately, combining this robust architecture with DDIM sampling yields a stable, highly efficient framework for low-data agricultural image synthesis. Code and models will be available at https://github.com/imagine-laboratory/Latent-Diffusion.
dc.description.wosFundingTextThis work was supported by the Flemish Government under the "Onderzoeksprogramma Artificiele Intelligentie (AI) Vlaanderen.
dc.identifier.doi10.1145/3795101.3814709
dc.identifier.isbn979-8-4007-2487-9
dc.identifier.urihttps://imec-publications.be/handle/20.500.12860/60811
dc.language.isoeng
dc.provenance.editstepusergreet.vanhoof@imec.be
dc.publisherASSOC COMPUTING MACHINERY
dc.source.beginpage1457
dc.source.conferenceGENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION - GECCO
dc.source.conferencedate2026-07-13
dc.source.conferencelocationSan José, Costa Rica
dc.source.endpage1464
dc.source.journalPROCEEDINGS OF THE 2026 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, GECCO 2026 COMPANION
dc.source.numberofpages8
dc.title

Peeling Back the Pixels: Stabilizing Latent Diffusion for Data-Efficient Generation of Pineapple Fields

dc.typeProceedings paper
dspace.entity.typePublication
imec.internal.crawledAt2026-11-14
imec.internal.sourcecrawler
imec.internal.wosCreatedAt2026-11-14
Files

Original bundle

Name:
3795101.3814709.pdf
Size:
1.48 MB
Format:
Adobe Portable Document Format
Description:
Published
Publication available in collections: