Daniel Panangian

RESEARCH ASSOCIATE @ GERMAN AEROSPACE CENTER Icon

00074-1025793139.png

I am part of AI4BuildingModeling Group headed by Dr. Ksenia Bittner. Our group specializes in 3D urban scene reconstruction using satellite imagery.

Unleashing the power of vision generative models for real-world applications is my main quest. I have been working mainly around inverse problems in imaging but not limited to image-to-image translation, topology-aware and graph-based approaches for building extraction.

Prior to research, I worked on pilot projects at industry-leading companies, developing language models to support internal processes. I earned both my Master’s and Bachelor’s degrees in Computer Science with a minor in Computational Linguistics from Ludwig Maximilian University Munich Icon.

publications

  1. twinworld_dsm.png
    Conditional Latent Diffusion Models for Digital Surface Model Generation★ NEW
    Debwashis Borman ,  Daniel Panangian ,  Pedro Achanccaray Diaz ,  Martin Eisemann , and  Ksenia Bittner
    In ECCVW 2026 TwinWorld , 2026
  2. roofdit.png
    RoofDiT: Diffusion Transformers for Roof Graph Synthesis and Reconstruction★ NEW
    Daniel Panangian , and  Ksenia Bittner
    In 14th IAPR Workshop on Pattern Recognition in Remote Sensing (PRRS), ICPRW 2026 , 2026
    Poster RoofDiT: Diffusion Transformers for Roof Graph Synthesis and Reconstruction poster
  3. roofvip.png
    RoofVIP Benchmark Dataset: 2D Roof Planar Polygons and Very High-Resolution Digital Orthophotos Pairs for Building Roof Reconstruction
    Chaikal Amrullah ,  Daniel Panangian ,  Guneet Mutreja ,  Youssef Abdelhedi , and  Ksenia Bittner
    ISPRS Annals of Photogrammetry, Remote Sensing, and Spatial Information Sciences , 2026
    Paper Website
  4. lod2former.png
    LoD2-Former: Multi-Modal Transformer-Based 3D Building Wireframe Reconstruction
    Youssef Abdelhedi ,  Daniel Panangian ,  Chaikal Amrullah ,  Houda Chaabouni-Chouayakh , and  Ksenia Bittner
    ISPRS Annals of Photogrammetry, Remote Sensing, and Spatial Information Sciences , 2026
    Paper
  5. PolyRoof: Precision Roof Polygonization in Urban Residential Building with Graph Neural Networks
    PolyRoof: Precision Roof Polygonization in Urban Residential Building with Graph Neural Networks
    Chaikal Amrullah ,  Daniel Panangian , and  Ksenia Bittner
    In Proceedings of the 17th International Joint Urban Remote Sensing Event (JURSE) , 2025
    Paper
  6. AIM2PC: Aerial Image to 3D Building Point Cloud Reconstruction
    AIM2PC: Aerial Image to 3D Building Point Cloud Reconstruction
    Soulaimene Turki ,  Daniel Panangian ,  Houda Chaabouni-Chouayakh , and  Ksenia Bittner
    ISPRS Annals of Photogrammetry, Remote Sensing, and Spatial Information Sciences , 2025
    Paper Code
  7. Dfilled: Repurposing Edge-Enhancing Diffusion for Guided DSM Void Filling
    Dfilled: Repurposing Edge-Enhancing Diffusion for Guided DSM Void Filling
    Daniel Panangian , and  Ksenia Bittner
    In IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) , 2025
    Paper Poster Dfilled: Repurposing Edge-Enhancing Diffusion for Guided DSM Void Filling poster
  8. Can Location Embeddings Enhance Super-Resolution of Satellite Imagery?
    Can Location Embeddings Enhance Super-Resolution of Satellite Imagery?
    Daniel Panangian , and  Ksenia Bittner
    In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , 2025
    Paper Poster Can Location Embeddings Enhance Super-Resolution of Satellite Imagery? poster
  9. SyntStereo2Real: Edge-aware GAN for remote sensing image translation while maintaining stereo constraints
    SyntStereo2Real: Edge-aware GAN for remote sensing image translation while maintaining stereo constraints
    Vasudha Venkatesan ,  Daniel Panangian ,  Mario F. Reyes , and  Ksenia Bittner
    In IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , 2024
    Paper YouTube Poster SyntStereo2Real: Edge-aware GAN for remote sensing image translation while maintaining stereo constraints poster
  10. Real-GDSR: Real-World Guided DSM Super-Resolution via Edge-Enhancing Residual Network
    Real-GDSR: Real-World Guided DSM Super-Resolution via Edge-Enhancing Residual Network
    Daniel Panangian , and  Ksenia Bittner
    ISPRS Annals of Photogrammetry, Remote Sensing, and Spatial Information Sciences , 2024
    Paper YouTube

news

Aug 10, 2026 Conditional Latent Diffusion Models for Digital Surface Model Generation is accepted to ECCVW 2026 TwinWorld
Jun 11, 2026 RoofDiT is accepted to PRRS 2026 (ICPRW 2026)
Dec 30, 2024 Dfilled is accepted to WACVW GeoCV
Oct 28, 2024 Can Location Embeddings Enhance Super-Resolution of Satellite Imagery? is accepted to WACV
Apr 05, 2024 SyntStereo2Real is accepted to CVPRW EarthVision
Mar 01, 2024 Real-GDSR is accepted to ISPRS TC II Symposium as oral presentation
Oct 10, 2023 Visited ICCV in Paris. First conference ever!