CoDEx: Combining Domain Expertise for Spatial Generalization in Satellite Image Analysis
Not on the map. Why?
Abstract (arXiv)
Global variations in terrain appearance raise a major challenge for satellite image analysis, leading to poor model performance when training on locations that differ from those encountered at test time. This remains true even with recent large global datasets. To address this challenge, we propose a novel domain-generalization framework for satellite images. Instead of trying to learn a single generalizable model, we train one expert model per training domain, while learning experts' similarity and encouraging similar experts to be consistent. A model selection module then identifies the most suitable experts for a given test sample and aggregates their predictions. Experiments on four datasets (DynamicEarthNet, MUDS, OSCD, and FMoW) demonstrate consistent gains over existing domain generalization and adaptation methods. Our code is publicly available at https://github.com/Abhishek19009/CoDEx.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 31 of its 46 references that are in the map. First posted on arXiv 28 Apr 2025.
Builds on
31 of 46 references.
References · 31
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
- 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,857
- 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,085
- 2019Invariant Risk Minimizationcited by 1,854
- 2020WILDS: A Benchmark of in-the-Wild Distribution Shiftscited by 1,375
- 2019Self-training with Noisy Student improves ImageNet classificationcited by 1,311
- 2016Deep CORAL: Correlation Alignment for Deep Domain Adaptationcited by 1,200
- 2020Out-of-Distribution Generalization via Risk Extrapolation (REx)cited by 742
- 2018ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentationcited by 623
- 2017Functional Map of the Worldcited by 338
- 2021Seasonal Contrast: Unsupervised Pre-Training from Uncurated Remote Sensing Datacited by 215
19 more
- 2020ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learningcited by 209
- 2020Universal Domain Adaptation through Self Supervisioncited by 172
- 2021Panoptic Segmentation of Satellite Image Time Series with Convolutional Temporal Attention Networkscited by 143
- 2019Domain Adaptation for Semantic Segmentation with Maximum Squares Losscited by 138
- 2018Urban Change Detection for Multispectral Earth Observation Using Convolutional Neural Networkscited by 110
- 2020Contrastive Learning for Label-Efficient Semantic Segmentationcited by 93
- 2021Extending the WILDS Benchmark for Unsupervised Adaptationcited by 86
- 2022DynamicEarthNet: Daily Multi-Spectral Satellite Dataset for Semantic Change Segmentationcited by 78
- 2021The Multi-Temporal Urban Development SpaceNet Datasetcited by 62
- 2021Multi-Modal Temporal Attention Models for Crop Mapping from Satellite Time Seriescited by 49
- 2022Domain Generalization using Pretrained Models without Fine-tuningcited by 29
- 2020StandardGAN: Multi-source Domain Adaptation for Semantic Segmentation of Very High Resolution Satellite Images by Data Standardizationcited by 23
- 2021TimeMatch: Unsupervised Cross-Region Adaptation by Temporal Shift Estimationcited by 21
- 2023Improving Domain Generalization with Domain Relationscited by 14
- 2023GeoMultiTaskNet: remote sensing unsupervised domain adaptation using geographical coordinatescited by 12
- 2023Pixel-wise Agricultural Image Time Series Classification: Comparisons and a Deformable Prototype-based Approachcited by 8
- 2024LFME: A Simple Framework for Learning from Multiple Experts in Domain Generalizationcited by 8
- 2024Satellite Image Time Series Semantic Change Detection: Novel Architecture and Analysis of Domain Shiftcited by 6
- 2021StyleAugment: Learning Texture De-biased Representations by Style Augmentation without Pre-defined Texturescited by 3
Built on by
Citing papers per year: 2026: 2.
2 papers.