Hello, I am Boaz
Portrait of Boaz Mwubahimana
Customize Theme

*Theme settings will be saved for
your next visit

What I Do

I am a Ph.D. candidate in Photogrammetry and Remote Sensing at LIESMARS, Wuhan University, working on GeoAI: deep learning semantic mapping for large-scale remote sensing, visual occlusion content completion, and forest information extraction.

More about me

I am a Ph.D. candidate at the State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), Wuhan University, working with Prof. Yan Jianguo. My research interests include deep learning semantic mapping for large-scale remote sensing applications, visual occlusion content completion, and forest information extraction. Alongside my career, I have over seven years of analysing imagery(remote sensing: satellite, drone and aiborne).

TOP EXPERTISE

GeoAI Research Engineer, Remote Sensing & Satellite image analysis: Download Resume

Be directed to my Certifications: View Certifications

  • Python (PyTorch)
  • Deep Learning (CNNs, ViTs)
  • Weak Supervision
  • Machine Learning
  • JavaScript
  • Photogrammetry, LiDAR & Point Cloud Analysis
  • Google Earth Engine
  • GIS software
  • Remote Sensing

📖 Academic Service

📝 Selected Publications

IEEE TGRS 2026 Graphical abstract

TagParaFormer: Cross-Hybrid Attention Learning Framework for Topology-Aware Road Network Extraction from Remote Sensing Imagery

IEEE Transactions on Geoscience and Remote Sensing, Early Access

Paper

Boaz Mwubahimana, Dingruibo Miao, Yan Jianguo, Le Ma, Remy Dukundane, Xiao Huang, et al.

IEEE TGRS 2026 Graphical abstract

FWDNNet: Cross-Heterogeneous Encoder Fusion via Feature-Level TensorDot Operations for Land-Cover Mapping

IEEE Transactions on Geoscience and Remote Sensing, vol. 64

Paper

Boaz Mwubahimana, Yan Jianguo, Dingruibo Miao, Swalpa Kumar Roy, Zhuohong Li, Le Ma, Clarisse Kagoyire, et al.

IEEE JSTARS 2026 Graphical abstract

Cross-Causal CarbonFormer: Rethinking Forest Carbon Density Estimation with a Causal Spatiotemporal Framework

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Open Access

Paper (Open Access)

Boaz Mwubahimana, Yan Jianguo, Dingruibo Miao, Zhuohong Li, Le Ma, Yves Ishimwe, et al.

IGARSS 2026 — Oral Graphical abstract

Top0CoverNet: Cross-Near 0 End-to-End Framework for Topologically Consistent Land Cover Mapping

IEEE International Geoscience and Remote Sensing Symposium, Washington, DC, USA — accepted for oral presentation; awarded conference travel grant

Boaz Mwubahimana, Dingruibo Miao, Yan Jianguo

IEEE TGRS 2025 Graphical abstract

C2FNet: Cross-Probabilistic Weak Supervision Learning for High-Resolution Land Cover Enhancement

IEEE Transactions on Geoscience and Remote Sensing, vol. 63

Paper

Boaz Mwubahimana, Yan Jianguo, Dingruibo Miao, Zhuohong Li, Haonan Guo, Le Ma, et al.

IJRS 2025 Graphical abstract

Vision Transformer-Based Feature Harmonization Network for Fine-Resolution Land Cover Mapping

International Journal of Remote Sensing, vol. 46, no. 10, pp. 3736–3769

Paper

Boaz Mwubahimana, Yan Jianguo, Maurice Mugabowindekwe, Xiao Huang, Elias Nyandwi, Joseph Tuyishimire, et al.

Preprint 2026 Graphical abstract

Cross Deep Neural Regression Learning for Aboveground Carbon Density Estimation from Multi-Temporal Multimodal Remote Sensing Data

Under review, 2026

Boaz Mwubahimana, Yan Jianguo, Dingruibo Miao, Maurice Mugabowindekwe, Swalpa Kumar Roy, et al.

Preprint 2025 Graphical abstract

TAG-Para: Hybrid Deep Attentions and Graphical Representations for Road Network Topological Error Correction

Available at SSRN: 5224096

Preprint

Boaz Mwubahimana, Yan Jianguo, Maurice Mugabowindekwe, Xiao Huang, Elias Nyandwi, et al.

🔥 News

💬 Talks and Oral Presentations

My featured projects

VCRP: Volcano Community Resilience Project |2024–2025

GeoAI for large-scale land cover mapping with multispectral imagery and deep learning, and forest information extraction using LiDAR — funded by the World Bank.

Ending Malaria with AI |2024–2025

Multispectral airborne sensing and AI to detect malaria hot-spot breeding sites from drone imagery — funded by the Bill & Melinda Gates Foundation.

Carbon Sink Estimation in Kigali |2024–2025

Deep neural regression framework for aboveground carbon density from multi-temporal Landsat imagery on Google Earth Engine (R² = 0.94, RMSE = 12.3 Mg C/ha).

sKwanda-V2 GeoAI Dataset |2024–2025

Large-scale training and ground-truth datasets from Rwanda and USA imagery, supporting open science and reproducible research.

View on GitHub
Omdena, Food security through ML

Omdena food security, Nakuru Kenya chapter.

Read More
Urban Growth Mapping in Kigali |2021

Global research poster: "Remote Sensing and Spatio-temporal Mapping of Urban Growth in Kigali, Rwanda. Case study of Kicukiro district".

Read More
E-learning platform |2021

Distance-learning platform where professionals become role models to young people and provide free mentorship sessions.

Read More
Research Paper |2021

"Physical Accessibility Analysis to Health Care Services in Rwanda, case study Musanze district", International Journal of Scientific & Engineering Research (IJSER), April 2021.

Read More
Kamagiri Physical Plan |2021

Rwanda Housing Authority — professional internship project including development of zoning plans.

Read More
Email: aiboaz1896@gmail.com

Get In Touch