Make Some Noise: Unsupervised Remote Sensing Change Detection Using Latent Space Perturbations

2026-02-23

AI Summary

This work addresses the limited generalization of unsupervised remote sensing change detection methods in rare or complex scenarios, which often stems from reliance on predefined change assumptions. To overcome this, we propose MaSoN, a framework that dynamically synthesizes diverse change samples in the latent feature space, aligning them with the target domain through statistical characteristics derived directly from the target data—without requiring handcrafted rules or external models. By integrating end-to-end training, feature-space perturbation, and data-driven statistical estimation, MaSoN naturally supports multimodal extensions (e.g., SAR) and significantly enhances generalization to unseen change types. Evaluated on five benchmark datasets, MaSoN achieves an average F1-score improvement of 14.1 percentage points over state-of-the-art methods, demonstrating superior performance.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Unsupervised change detection (UCD) in remote sensing aims to localise semantic changes between two images of the same region without relying on labelled data during training. Most recent approaches rely either on frozen foundation models in a training-free manner or on training with synthetic changes generated in pixel space. Both strategies inherently rely on predefined assumptions about change types, typically introduced through handcrafted rules, external datasets, or auxiliary generative models. Due to these assumptions, such methods fail to generalise beyond a few change types, limiting their real-world usage, especially in rare or complex scenarios. To address this, we propose MaSoN (Make Some Noise), an end-to-end UCD framework that synthesises diverse changes directly in the latent feature space during training. It generates changes that are dynamically estimated using feature statistics of target data, enabling diverse yet data-driven variation aligned with the target domain. It also easily extends to new modalities, such as SAR. MaSoN generalises strongly across diverse change types and achieves state-of-the-art performance on five benchmarks, improving the average F1 score by 14.1 percentage points. Project page: https://blaz-r.github.io/mason_ucd

Problem

• Unsupervised change detection
• Remote sensing
• Generalization
• Latent space
• Synthetic changes

Innovation

• latent space perturbations
• unsupervised change detection
• feature space synthesis
• remote sensing
• domain-adaptive generation