Semantic change detection learns new land categories incrementally
HyperFlow: Class-Incremental Semantic Change Detection via Hyperbolic Flow
Computer Vision and Pattern Recognition
Summary
Satellite images show how land use changes over time, like forests turning into farms. Detecting these changes is tricky because new types of changes can appear that the model hasn't seen before. The authors created HyperFlow, which lets a computer learn new kinds of land changes step-by-step, using a special math space that understands how categories relate to each other. They also made a big new dataset to test this approach. Their method improves how well changes and land types are detected as new categories are added.
What this means in practice
- •For environmental monitoring teams: Detect new land-cover changes progressively in satellite images to improve environmental assessments over time.
- •For urban planning departments: Identify and track emerging types of land use changes in city areas as categories evolve incrementally.
Authors
Manon Béchaz, Léo Buecher, Emanuele Dalsasso, Ciprian Tomoiaga, Patrick Jayet, Marcin Detyniecki, Laetitia Chapel, Lucas Drumetz, Nicolas Courty, Devis Tuia
Abstract
Semantic change detection in remote sensing aims to identify and characterize land-cover transitions observed in pairs of satellite or aerial images. Existing methods assume a fixed set of change categories, limiting their applicability in real-world scenarios where changes of interest may vary, for example across geographies or over time. In this work, we tackle the problem of class-incremental semantic change detection, where a model progressively learns to detect new types of semantic change. To do so, we exploit the hierarchical structure of land-cover semantics, where newly introduced categories correspond to fine-grained refinements of existing ones, and represent said semantic categories in hyperbolic space, which is particularly well suited to represent such hierarchical relationships. To capture semantic changes, we introduce HyperFlow, a method relying on Busemann prototypes, modeling change as an anomaly in a learned semantic flow within the hyperbolic embedding space induced by segmentation. HyperFlow couples change detection to class-incremental semantic segmentation, enabling class-incremental semantic change detection. We also present HiChange, a new large-scale, high resolution remote sensing dataset with detailed, fine-grained hierarchical labels. Experimental results demonstrate that HyperFlow achieves both strong change detection and semantic segmentation performance across incremental steps. In particular, we show that progressively refining semantic categories, rather than considering them all directly, yields improved change detection, highlighting the benefits of incremental learning for semantic change detection.