thesis

Are you interested in a research-oriented thesis in Computer Vision, combining geometry and learning?

It is good to have an end to journey toward; but it is the journey that matters, in the end. Hemingway

A thesis has a destination, but it is also a process: learning how to think rigorously, how to model a problem, and how to build something that works while understanding why it works.
I am particularly interested in Geometry-Driven Vision: designing practical vision systems that combine geometric reasoning with modern learning techniques.

Available thesis topics:

  • :mortar_board: The complete and updated list of thesis opportunities in Computer Vision, Deep Learning, Image Processing is available.

  • :mag: Within these broader research directions, I am currently looking for motivated students interested in:

    :sailboat: Direct 3D Curve Reconstruction for Sports Applications
    Analytic reconstruction of 3D curves directly from multi-view images, with applications to high-performance sailing and shape optimization.

    :skull: Skeleton-Based Multi-View Calibration
    Privacy-preserving multi-camera calibration directly from 2D pose estimation outputs, embedding articulated constraints into a robust geometric optimization framework.

    :package: Geometry-Aware 3D Reconstruction in Structured Environments
    Joint recovery of scene geometry and camera poses in warehouse environments by combining vanishing-point reasoning, Manhattan priors, and learned components.

  • :envelope: If you would like to be updated on these or other upcoming opportunities, feel free to drop me an email.
  • :star: You can visit our Thesis Hall of Fame to get an idea of past successful theses (some of which have been published at workshops, conferences, and journals).

Other Topics of Interest

Beyond the specific theses listed above, my research interests within Geometry-Driven Vision include:

  • Hybrid deep-geometric pipelines for 3D vision and structured scene understanding
  • Robust multi-view geometry and camera calibration
  • Learning inside geometric optimization and model fitting
  • Robust estimation and primitive decomposition for real-world geometric problems
  • Geometry-aware self-supervision and learning with limited labels
  • Pattern recognition in structured domains, including clustering, anomaly detection, and explainability

Background and Preparation

Do not worry if some background is missing, small gaps can be bridged at the beginning of a thesis, and we will provide references when needed. A solid starting point typically includes courses such as IACV, AN2DL, or MMMIP, or equivalent preparation in multi-view geometry, deep learning, and image processing. Nobody knows everything before starting, this is not an exam session!
By the time you graduate, however, you should master every aspect of your work. We also organize a series of Thesis-How-To meetings to provide structured guidance throughout the thesis.

Theses supervised

You can have a look at the theses I supervised to get a sense of the possible research directions:

  • Alessandro Ardenghi, Teaching CAMs to Remember: a Memory-Enhanced approach to Weakly Supervised Video Semantic . MSc in Artificial Intelligence, December 2025, Bocconi University.
  • Francesco Santambrogio, A Frequency Loss Driven Framework for Respiration Monitoring Using Depth-Sensing Cameras. MSc in Computer Science and Engineering. October 2025, PoliMi.
  • Francesco Buccoliero, Template Matching for Logos Exploiting Geometry and Color Priors. MSc in Computer Science and Engineering. October 2025, PoliMi.
  • Stefano Arcaro, Single-view 3D Reconstruction from Planar Templates. MSc in Computer Science and Engineering. October 2025, PoliMi.
  • Enrico Barbieri, Pose and camera calibration from cylinders via a silhouette based homotopy continuation. MSc in Computer Science and Engineering. July 2025, PoliMi.
  • Caterina Giardi, Heart Rate estimation from videos: a robust approach for rPPG signals analysis. MSc in Computer Science and Engineering. July 2025, PoliMi.
  • Mattro Forlivesi, Critical Configurations in Two-View Reconstruction: A Degeneracy Test for Fundamental Matrix Estimation. MSc in Computer Science and Engineering. April 2025, PoliMi.
  • Andrea Bertogalli, Multi-modal sensing for autonomous driving: a unified calibration framework for LiDAR, RGB and Event cameras. MSc in Computer Science and Engineering. April 2025, PoliMi.
  • Yassine Ouhadi, Explainable Agglomerative Clustering Tree: ExACt. MSc in Computer Science and Engineering. April 2025, PoliMi.
  • Andrea Ferraris, Geometric-Aware Local Optimization for Primitive Fitting. MSc in Computer Science and Engineering. December 2024, PoliMi.
  • Andrea Naclerio, Vision-Based Heart Rate Monitoring: Towards Adaptive Dynamic ROI Selection. MSc in Biomedical Engineering. December 2024, PoliMi.
  • Erica Manfrin, Structure-based anomaly detection on 3D Point Clouds. MSc in Mathematical Engineering. October 2024, PoliMi.
  • Andrea Bonatti, Liquid Lens Calibration. MSc in Computer Science and Engineering. July 2024, PoliMi.
  • Barini Matteo, Image rectification with a known template. MSc in Mathematical Engineering. March 2024, PoliMi.
  • Simone Colombara, Computing correspondences for Partial Shape Matching via Functional Maps. MSc in Mathematical Engineering. December 2023, PoliMi.
  • Francesco Azzoni, Deep Monocular Autocalibration of Radially Symmetric Wide-Angle Cameras. MSc in Computer Science and Engineering. October 2023, PoliMi.
  • Leonardo Perelli, Trifocal Tensor Estimation for n-view Deep Structure-from-Motion. MSc in Mathematical Engineering. October 2022, PoliMi.
  • Lucia Gioria, Improving Explainable Clustering via Probabilistic Modeling. MSc in Mathematical Engineering. October 2022, PoliMi.
  • Valentina Sgarbossa, Visual Localization in presence of match scarcity. MSc in Mathematical Engineering. October 2022, PoliMi.
  • Danilo Catone, Anomaly detection via learned models and preference analysis. MSc in Computer Science and Engineering. October 2022, PoliMi.
  • Giuseppe Bertolini, Towards a practical, fully automated joint calibration pipeline for X-Ray-RGB medical imaging systems. MSc in Computer Science and Engineering. April 2022. PoliMi.
  • Filippo Galli, Camera Pose Estimation for planar configurations. MSc in GeoInformatics. April 2022. PoliMi.
  • Enrico Ruggiano, Detezione di Linee in Immagini Distorte tramite Analisi delle Preferenze, MSc in Computer Science and Engineering. April 2022. PoliMi.
  • Diana Isaeva, Robust homography estimation for image matching. MSc in Mathematical Engineering. December 2021, PoliMi.
  • Andrea Porfiri Dal Cin, Synchronization on Group-labelled Multigraphs. MSc in Computer Science and Engineering. April 2021, PoliMi.
  • Antonino Maria Rizzo , Semantic aware Sampling for Robust Multi-model Fitting. MSc in Computer Science and Engineering. April 2021, PoliMi.
  • William Bonvini, Unsupervised Learning for Multi-Model Consensus Maximization. MSc in Computer Science and Engineering. April 2021, PoliMi.
  • Simone Francavilla, Image Mosaicing: An approach based on Synchronization and Game Theory. MSc in Computer Science and Engineering. April 2021, PoliMi.