SLUM-i: Semi-supervised Learning for Urban Mapping of Informal Settlements and Data Quality Benchmarking
arXiv:2602.04525v2 Announce Type: replace-cross Abstract: Rapid urban expansion has fueled the growth of informal settlements in major cities of low- and middle-income countries, with Lahore and Karachi in Pakistan and Mumbai in India serving as prominent examples. However, large-scale mapping of these settlements is severely constrained not only by the scarcity of annotations but by inherent data quality challenges, specifically high spectral ambiguity between formal and informal structures and significant annotation noise. We address this by introducing a benchmark dataset for Lahore, constr
The paper addresses a critical need for efficient and accurate urban mapping in rapidly expanding low- and middle-income country cities, driven by the increasing availability of satellite data and advancements in semi-supervised learning techniques.
Accurate mapping of informal settlements is crucial for urban planning, resource allocation, and addressing humanitarian challenges, and this development signals progress in using AI for such complex geospatial tasks.
The explicit benchmarking of data quality challenges and the introduction of a new dataset for Lahore set a new standard for research in this niche, potentially improving the reliability of AI applications in urban development.
- · Urban planners in LMICs
- · NGOs focused on humanitarian aid
- · Geospatial AI companies
- · Local governments in rapidly urbanizing areas
- · Traditional manual surveying methods
- · Urban areas lacking up-to-date mapping data
Improved understanding and management of informal settlements, leading to better resource allocation.
Reduced social inequalities and more effective provision of services in previously unmapped areas.
Enhanced data-driven policy-making in LMICs, potentially attracting investment and improving living standards.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI