Pedro Hermosilla

Assistant Professor

prof_pic.jpg

I am Assistant Professor in AI for Visual Computing at the Computer Vision Lab, at TU Wien, Austria. Before joining TU Wien, I did my PostDoc at the Viscom group headed by Prof. Timo Ropinski at Ulm University while working in close collaboration with Prof. Tobias Ritschel from University College London. I obtained my PhD at the Polytechnical University of Catalonia, under the supervision of Prof. Pere-Pau Vazquez and Prof. Alvar Vinacua.

My research is focused on developing machine learning technologies for 3D and unstructured data, with a special interest in point clouds, graphs, and implicit representations. During my research, I applied these technologies to solve different problems in the fields of Computer Vision, Computer Graphics, and Bioinformatics.

News

Feb 28, 2024 We have two papers accepted at CVPR. Congratulations to Sebastian Koch and Leon Sick!
Feb 20, 2024 Congratulations to Aron S. Kovacs for getting his work accepted at Eurographics 2024.
Jan 16, 2024 Our work on weakly-supervised virus capsid detection from electron microscopy images has been accepted to ICLR 2024. Congratulations to Hannah Kniesel!
Dec 28, 2023 The paper of Philipp Erle on surface reconstruction from point clouds has been accepted on Computer Graphics Forum.
Oct 21, 2023 The works of Lisa Weijler and Sebastian Koch have been accepted into WACV 2024.

Selected publications

  1. Open3DSG: Open-Vocabulary 3D Scene Graphs from Point Clouds with Queryable Objects and Open-Set Relationships
    S. Koch, N. Vaskevicius, M. Colosi,  P. Hermosilla, and T. Ropinski
    Conference on Computer Vision and Pattern Recognition (CVPR) 2024
  2. Unsupervised Semantic Segmentation Through Depth-Guided Feature Correlation and Sampling
    L. Sick, D. Engel,  P. Hermosilla, and T. Ropinski
    Conference on Computer Vision and Pattern Recognition (CVPR) 2024
  3. Weakly Supervised Virus Capsid Detection with Image-Level Annotations in Electron Microscopy Images
    H. Kniesel, L. Sick, T. Payer, T. Bergner, K. S. Devan, C. Read, P. Walther, T. Ropinski, and P. Hermosilla
    International Conference on Learning Representations (ICLR) 2024
  4. Lang3DSG: Language-based contrastive pre-training for 3D scene graph prediction
    S. Koch,  P. Hermosilla, N. Vaskevicius, M. Colosi, and T. Ropinski
    International Conference on 3D Vision (3DV) 2024
  5. Surface-aware Mesh Texture Synthesis with Pre-trained 2D CNNs
    A. S. Kovacs,  P. Hermosilla, and R. G. Raidou
    Computer Graphics Forum (Proc. Eurographics) 2024
  6. SGRec3D: Self-Supervised 3D Scene Graph Learning via Object-Level Scene Reconstruction
    S. Koch,  P. Hermosilla, N. Vaskevicius, M. Colosi, and T. Ropinski
    Winter Conference on Applications of Computer Vision (WACV) 2024
  7. FATE: Feature-Agnostic Transformer-Based Encoder for Learning Generalized Embedding Spaces in Flow Cytometry Data
    L. Weijler, F. Kowarsch, M. Reiter,  P. Hermosilla, M. Maurer-Granofszky, and M. Dworzak
    Winter Conference on Applications of Computer Vision (WACV) 2024
  8. PPSurf: Combining Patches and Point Convolutions for Detailed Surface Reconstruction
    P. Erler, L. Fuentes-Perez,  P. Hermosilla, P. Guerrero, R. Pajarola, and M. Wimmer
    Computer Graphics Forum 2024
  9. Weakly-Supervised Optical Flow Estimation for Time-of-Flight
    M. Schelling,  P. Hermosilla, and T. Ropinski
    Winter Conference on Applications of Computer Vision (WACV) 2023
  10. Variance-Aware Weight Initialization for Point Convolutional Neural Networks
    P. Hermosilla, M. Schelling, T. Ritschel, and T. Ropinski
    European Conference on Computer Vision (ECCV) 2022
  11. Clean Implicit 3D Structure from Noisy 2D STEM Images
    H. Kniesel, T. Ropinski, T. Bergner, K. Shaga Devan, C. Read, P. Walther, T. Ritschel, and P. Hermosilla
    Conference on Computer Vision and Pattern Recognition (CVPR) 2022
  12. RADU: Ray-Aligned Depth Update Convolutions for ToF Data Denoising
    M. Schelling,  P. Hermosilla, and T. Ropinski
    Conference on Computer Vision and Pattern Recognition (CVPR) 2022
  13. Gaussian Mixture Convolution Networks
    A. Celarek,  P. Hermosilla, B. Kerbl, T. Ropinski, and M. Wimmer
    International Conference on Learning Representations (ICLR) 2022
  14. Contrastive Representation Learning for 3D Protein Structures
    P. Hermosilla, and T. Ropinski
    Pre-print 2022
  15. Intrinsic-Extrinsic Convolution and Pooling for Learning on 3D Protein Structures
    P. Hermosilla, M. Schaefer, M. Lang, G. Fackelmann, P.-P. Vazquez, B. Kozlikova, M. Krone, T. Ritschel, and T. Ropinski
    International Conference on Learning Representations (ICLR) 2021
  16. Enabling Viewpoint Learning through Dynamic Label Generation
    M. Schelling,  P. Hermosilla, and T. Ropinski
    Computer Graphics Forum (Proc. Eurographics) 2021
  17. Total Denoising: Unsupervised Learning of 3D Point Cloud Cleaning
    P. Hermosilla, T. Ritschel, and T. Ropinski
    International Conference on Computer Vision (ICCV) 2019
  18. Deep-learning the Latent Space of Light Transport
    P. Hermosilla, S. Maisch, T. Ritschel, and T. Ropinski
    Computer Graphics Forum (Proc. EGSR) 2019
  19. Monte Carlo Convolution for Learning on Non-Uniformly Sampled Point Clouds
    P. Hermosilla, T. Ritschel, P.-P. Vazquez, A. Vinacua, and T. Ropinski
    ACM Transactions on Graphics (Proc. SIGGRAPH Asia) 2018