DODA: Diffusion for Object-detection Domain Adaptation in Agriculture

ICLR 2025 Conference Submission872 Authors

Keywords: diffusion model, object detection, domain adaptation

TL;DR: Domain-Specific Detection Data Generation for Diverse Agricultural Scenarios Via Multi-conditional Diffusion

Abstract: Object detection has wide applications in agriculture, but the trained models often struggle to generalize across diverse agricultural environments. To address this challenge, we propose DODA (\underline{D}iffusion for \underline{O}bject-detection \underline{D}omain Adaptation in \underline{A}griculture), a unified framework that leverages diffusion models to generate domain-specific detection data for multiple agricultural scenarios. DODA incorporates external domain embeddings and an improved layout-to-image (L2I) approach, allowing it to generate high-quality detection data for new domains without additional training. We demonstrate DODA's effectiveness on the Global Wheat Head Detection dataset, where fine-tuning detectors on DODA-generated data yields significant improvements across multiple domains (maximum +15.6 AP). DODA provides a simple yet powerful approach to adapt object detectors to diverse agricultural scenarios, lowering barriers for more plant breeders growers to use detection in their personalized environments.

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 872