# Unsupervised Domain Adaptation For Plant Organ Counting

arXiv: https://arxiv.org/abs/2009.01081 · HTML: https://mlsanity.com/paper/2009.01081/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Sep 2020.

## Built on by (most cited first, 7 of 7)

- [Label-Efficient Learning in Agriculture: A Comprehensive Review](https://mlsanity.com/paper/2305.14691/) (2023)
- [Semi-supervised dry herbage mass estimation using automatic data and synthetic images](https://mlsanity.com/paper/2110.13719/) (2021)
- [AlphaGarden: Learning to Autonomously Tend a Polyculture Garden](https://mlsanity.com/paper/2111.06014/) (2021)
- [Automated Pruning of Polyculture Plants](https://mlsanity.com/paper/2208.10472/) (2022)
- [On Domain-Specific Pre-Training for Effective Semantic Perception in Agricultural Robotics](https://mlsanity.com/paper/2303.12499/) (2023)
- [Counting Objects by Diffused Index: geometry-free and training-free approach](https://mlsanity.com/paper/2110.08365/) (2021)
- [Wheat Head Counting by Estimating a Density Map with Convolutional Neural Networks](https://mlsanity.com/paper/2303.10542/) (2023)

## Abstract (arXiv)

Supervised learning is often used to count objects in images, but for counting small, densely located objects, the required image annotations are burdensome to collect. Counting plant organs for image-based plant phenotyping falls within this category. Object counting in plant images is further challenged by having plant image datasets with significant domain shift due to different experimental conditions, e.g. applying an annotated dataset of indoor plant images for use on outdoor images, or on a different plant species. In this paper, we propose a domain-adversarial learning approach for domain adaptation of density map estimation for the purposes of object counting. The approach does not assume perfectly aligned distributions between the source and target datasets, which makes it more broadly applicable within general object counting and plant organ counting tasks. Evaluation on two diverse object counting tasks (wheat spikelets, leaves) demonstrates consistent performance on the target datasets across different classes of domain shift: from indoor-to-outdoor images and from species-to-species adaptation.