![[2507.22554] DeepC4: Deep Conditional Census-Constrained Clustering for Large-scale Multitask Spatial Disaggregation of Urban Morphology](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2507.22554] DeepC4: Deep Conditional Census-Constrained Clustering for Large-scale Multitask Spatial Disaggregation of Urban Morphology
Summary
The paper presents DeepC4, a novel deep learning approach for spatial disaggregation of urban morphology, enhancing mapping quality using census data and satellite imagery.
Why It Matters
This research addresses significant challenges in urban mapping, particularly in developing economies, by integrating local census statistics with advanced deep learning techniques. As global frameworks approach their 2030 goals, accurate urban morphology mapping is crucial for sustainable development and disaster risk reduction.
Key Takeaways
- DeepC4 improves urban morphology mapping by incorporating census data.
- The method addresses local discrepancies in existing mapping techniques.
- It enhances explainability and interpretability in large-scale mapping efforts.
- The approach is particularly beneficial for developing economies.
- DeepC4 supports global sustainable development goals through better data utilization.
Computer Science > Machine Learning arXiv:2507.22554 (cs) [Submitted on 30 Jul 2025 (v1), last revised 15 Feb 2026 (this version, v2)] Title:DeepC4: Deep Conditional Census-Constrained Clustering for Large-scale Multitask Spatial Disaggregation of Urban Morphology Authors:Joshua Dimasaka, Christian Geiß, Emily So View a PDF of the paper titled DeepC4: Deep Conditional Census-Constrained Clustering for Large-scale Multitask Spatial Disaggregation of Urban Morphology, by Joshua Dimasaka and 2 other authors View PDF HTML (experimental) Abstract:To understand our global progress for sustainable development and disaster risk reduction in many developing economies, two recent major initiatives - the Uniform African Exposure Dataset of the Global Earthquake Model (GEM) Foundation and the Modelling Exposure through Earth Observation Routines (METEOR) Project - implemented classical spatial disaggregation techniques to generate large-scale mapping of urban morphology using the information from various satellite imagery and its derivatives, geospatial datasets of the built environment, and subnational census statistics. However, the local discrepancy with well-validated census statistics and the propagated model uncertainties remain a challenge in such coarse-to-fine-grained mapping problems, specifically constrained by weak and conditional label supervision. Therefore, we present Deep Conditional Census-Constrained Clustering (DeepC4), a novel deep learning-based spatial disaggrega...