GeoFormer: A Swin Transformer-Based Framework for Scene-Level Building Height and Footprint Estimation from Sentinel Imagery

Zenodo (CERN European Organization for Nuclear Research) 2026 Dataset Open access

Read full text
Authors: Jinzhen Han
DOI: 10.5281/zenodo.18573862
Cited by: 0

Abstract

This repository contains the preprocessed dataset (.npz format) used for training and evaluating the GeoFormer framework — a Swin Transformer-based multi-task model for joint building height (BH) and building footprint (BF) estimation at 100 m resolution. ⚠️ Status: Coming Soon The dataset files are being prepared and will be uploaded shortly. Please check back later or follow this record for updates. Related resources: Paper: (under review)

Fields of Study & Subjects

Automated Road and Building Extraction Advanced Neural Network Applications 3D Modeling in Geospatial Applications

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
10 February 2026
Language
EN
Document type
Dataset