Rethinking Metrics and Diffusion Architecture for 3D Point Cloud Generation
Bastico, Matteo, et al. Rethinking Metrics and Diffusion Architecture for 3D Point Cloud Generation. 2025, https://arxiv.org/abs/2511.05308.
Education
- PhD in Deep Learning and 3D Vision in Healthcare, Doctoral program AI4TheSciences - Horizon 2020 -Marie Skłodowska-Curie Actions-COFUND, Mines Paris - PSL Univeristy, France, 2025
- Led the design and development of large-scale deep-learning architectures for 3D vision across multiple modalities (images and point clouds) and tasks (segmentation, landmarking, synthesis), with research published at CVPR, ICCVW, and 3DV.
- Experienced in cross-functional collaboration and bridging research innovation with practical healthcare challenges. In addition, I trained and supervised two master students.
- Thesis “Advancing 3D Vision and Deep Learning for Shape Analysis and Synthesis: Application to Cross-Species Knee Joint Biomechanical Modeling”.
- MSc in ICT for Internet and Multimedia cum laude, Università degli Studi di Padova, Italy, 2021
- Thesis “Simultaneous Depth Completion and 3D Object Detection via Deep Learning for Scene Reconstruction in Autonomous Driving Scenarios”.
- MSc in Telecommunication Engineering, Universidad Politecnica de Madrid, Spain, 2021
- BSc in Information Engineering, Università degli Studi di Padova, Italy, 2018
Work experience
- 2021 - 2022: Deep Learning Research Scientist @ Universidad Politécnica de Madrid, Spain
- Contributed to EU Horizon 2020 projects, such as GenoMed4All and PROCare4Life, coordinating work across technical teams and developing supervised deep learning models in PyTorch for DNA, medical imaging, and 3D skeletal data.
- Delivered end-to-end pipelines and co-authored 3 peer-reviewed publications.
- 2018 - 2019: ICT Help Desk Agent @ IL GAZZETTINO, Venice, Italy
- 2014 - 2015: Front-End Developer @ Valore4IT, Venice, Italy
Scholarships
- 2022 - 2025: Doctoral program AI4TheSciences - Horizon 2020 -Marie Skłodowska-Curie Actions-COFUND @ PSL University, France
- 2019 - 2021: Double Master’s Degree program @ University of Padua (Italy) & Universidad Politécnica de Madrid (Spain)
Skills
- Transformers, CNNs
- Generative Diffusion Models
- Medical Imaging
- Semantic Segmentation
- 3D Point Clouds
- Python, PyTorch
- Scalable Training on HPC
- Git, GitHub
Publications
Coupled Laplacian Eigenmaps for Locally-Aware 3D Rigid Point Cloud Matching
M. Bastico, E. Decencière, L. Corté, Y. Tillier and D. Ryckelynck, "Coupled Laplacian Eigenmaps for Locally-Aware 3D Rigid Point Cloud Matching," 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, 2024, pp. 3447-3458, doi: 10.1109/CVPR52733.2024.00331.
A Simple and Robust Framework for Cross-Modality Medical Image Segmentation applied to Vision Transformers
M. Bastico, D. Ryckelynck, L. Corté, Y. Tillier and E. Decencière, "A Simple and Robust Framework for Cross-Modality Medical Image Segmentation applied to Vision Transformers," 2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), Paris, France, 2023, pp. 4130-4140, doi: 10.1109/ICCVW60793.2023.00446.
DrOGA: An Artificial Intelligence Solution for Driver-Status Prediction of Genomics Mutations in Precision Cancer Medicine
M. Bastico, A. Fernández-García, A. Belmonte-Hernández and S. U. Mayoral, "DrOGA: An Artificial Intelligence Solution for Driver-Status Prediction of Genomics Mutations in Precision Cancer Medicine," in IEEE Access, vol. 11, pp. 37378-37391, 2023, doi: 10.1109/ACCESS.2023.3266983.
Simultaneous Real-Time Human Fall Detection and Reidentification Based on Multisensors Data
M. Bastico, V.R. Bejerano, and A. Belmonte-Hernández. 2022. Simultaneous Real-Time Human Fall Detection and Reidentification Based on Multisensors Data. In Proceedings of the 15th International Conference on PErvasive Technologies Related to Assistive Environments (PETRA 22). Association for Computing Machinery, New York, NY, USA, 365–370. https://doi.org/10.1145/3529190.3534728
Continuous Person Identification and Tracking in Healthcare by Integrating Accelerometer Data and Deep Learning Filled 3D Skeletons
M. Bastico, A. Belmonte-Hernández and F. Á. García, "Continuous Person Identification and Tracking in Healthcare by Integrating Accelerometer Data and Deep Learning Filled 3D Skeletons," in IEEE Sensors Journal, vol. 22, no. 15, pp. 15402-15409, 1 Aug.1, 2022, doi: 10.1109/JSEN.2022.3186499.
Simultaneous Depth Completion and 3D Object Detection via Deep Learning for Scene Reconstruction in Autonomous Driving Scenarios
M. Bastico, “Simultaneous Depth Completion and 3D Object Detection via Deep Learning for Scene Reconstruction in Autonomous Driving Scenarios,” MSc dissertation, Universidad Politecnica de Madrid, Spain, 2021.
Talks
Rethinking Metrics for Evaluating 3D Point Cloud Generation
Talk at French German Workshop, Kaiserslautern, Fraunhofer ITWM, Germany
Multi-modal Medical Image Segmentation
Talk at Seminar on Image Segmentation in Mechanics, ENS Paris-Saclay, France
Laplacian Eigenmaps for Anatomical Landmarks Detection and Knee Kinematics Axis Identification
Talk at International Conference on Biomedical Technology (ICBT), Hannover, Germany
Teaching
An Introduction to (Vision) Transformers
PSL week Machine Learning for Physics and Engineering, Mines Paris - PSL University, 2025
Image Segmentation: From MLP to ViT
PSL week Machine Learning for Science and Engineering, Mines Paris - PSL University, 2024
Deep representation Learning of Radiomics Features
Supervised Master’s Students during Trimestre recherche DIMA, Mines Paris - PSL University & Universidad Politècnica de Madrid, 2024
Transfer Learning for Anatomical Landmark Prediction in 3D Point Clouds
Supervised Master’s Students during Trimestre recherche DIMA, Mines Paris - PSL University, 2024
Spoken Languages
- Italian (First Language)
- English (C1)
- Spanish (B2-C1)
- French (B2-C1)