Doctoral School Projects
Master Thesis or Doctoral School Projects
- Spacecraft Dataset Development for Unseen and Occluded Targets
- Exploring the Future of Classical Music
- Quantized Neural Networks for Space Applications
- Monocular 3D Pose Estimation with Uncertainty Estimation for Handling Occlusion
- Improving human body models via neural ODE flow based implicit layer
- Self-supervision signals for video-based human mesh recovery
- Human Pose Estimation for Martial Arts
- Iterative Human Pose Refinement via Energy Ascent
- Learning silhouette appearance to improve multi-people tracking [master]
- Heritage in the digital age
- Deep Learning for Nuclear Fusion
- Promoting Connectivity of Linear Structures in 3D Microscopy Images
- Modeling People and their Clothes in Crowded Scenes
- Multi-task Active Learning
- Physically Constrained Deep Networks
For further project offers, please contact members of the CVLAB directly.
Semester Projects (Bachelor and Master)
Most of the offered semester projects can be rephrased for a thesis and vice versa. Please contact us directly.
Administrative
Semester Projects (Bachelor and Master)
SIN and SSC students do one semester project during their Bachelor studies and one semester project during their Master studies.
Semester projects can be done in groups of two students.
Semester projects are worth 8 credits for Bachelor and 12 credits for Master.
Students must have the approval of the Professor in charge of the laboratory before registering for the given project.
Oral defense: within two weeks of the hand-in date.
Master Thesis Projects
Master Thesis Projects are started once the complete master program is finished and all the credits have been obtained.
Projects for SSC and SIN students should last 4 months at the EPFL or 6 months in the industry or in another University.
Master Thesis Projects must be done individually.
Master Thesis Projects are worth 30 credits.
Students must have the approval of the Professor in charge of the laboratory before registering for the given project.
Additional information

Five instances of Liver extracted from MRI data. The black dot indicates a pseudo key point.In this project, we plan to explore methods to match/align shapes extracted from biomedical images, based on shape and texture-based features. By doing so, we will be able to analyze the variations of these structures over time and compare them (…)

DescriptionPrecise human body mesh recovery (or HMR) is a long-standing research topic due to its crucial role for human motion understanding. Recent works proposed very effective solutions for single-frame inputs (as original HMR [1] or more data-efficient [2]) that give the prediction for a body-centered image as well as for body-centered videos (e.g., TCMR [3]).When (…)

DescriptionRecent parametric human body models, such as SMPL [1] or GHUM [2] provide rich and natural yet very sparse prior for realistic body modelling. All parameters have semantic meaning and allow to render diverse set of human bodies in the form of triangular meshes with known topology. Usually, the set of parameters consists of disentangled (…)
DescriptionModern CNN-based human pose estimators are constructed as regression architectures with human pose prediction as output. This formulation does not naturally involve the confidence of the estimation. Another possible solution is to regress the pose refinement directly via the CNN [2], which gives slight improvement over the ordinary regression-based techniques. However, it still suffers from (…)

Invision system tracking and geo-localizing passengers in a train station using 8 cameras. p { line-height: 115%; text-align: left; orphans: 2; widows: 2; margin-bottom: 0.25cm; direction: ltr; background: transparent }a:link { color: #000080; so-language: zxx; text-decoration: underline } Invision AI develops a multi-camera system to track people and cars. The current system sometimes mixes up persons (…)

Deep learning based approaches can be used for a wide range of space applications such as on-board data processing for observation satellite and collision prevention, spacecraft rendezvous, etc. Unfortunately, the deep learning models are very computational intensive and require huge amount of resources and power consumption. In recent years, some techniques like quantization, pruning and (…)

If traditional services (loans, consultation, etc.) remain at the heart of the activities of the Cantonal University Library of the Canton of Fribourg (BCU), enhancing and sharing the digital heritage presents new challenges for libraries. In particular, how to rethink the representation of heritage in the digital era? How to create a relationship between physical (…)

An active research field is the prediction of physics mechanisms degrading plasma confinement and performance, eventually leading to disruptions on Tokamaks. Disruptions are the ultimate consequence of highly coupled non-linear plasma physics processes, resulting in an abrupt loss of plasma current and confinement inducing huge electromagnetic forces and thermal loads on Plasma Facing Components and (…)

As in many areas of computer vision, deep networks now deliver state-of-the-art results for delineation tasks, such as finding axons and dendrites in 3D light microscopy images. Most of the existing approaches rely on convolutional networks to extract from images binary masks denoting which voxels belong to neurites and which do not. Unfortunately, they do (…)

Fig 1: Common mistakes made by U-Net (inside orange squares)Current state-of-the art volumetric image segmentation approaches achieve impressive performances. But still, they are not 100% correct and make mistakes as you can see in Figure 1. Therefore we would like to know when a network makes a mistake, so that a human or another model (…)

While modeling people wearing tight-fitting clothes is fast becoming a mature field, handling subjects wearing looser garments remains an open problem when it is to be done in everyday settings where people may hide each other, precise outlines are hard to estimate, and shadows often complicate matters. Our goal is therefore to develop robust and (…)

DescriptionDifferent visual tasks are often strongly and obviously correlated. For instance, having surface normals simplifies estimating the depth of an image, knowing segmentation could help detect objects, etc. Our intuition implies the existence of a special structure among visual tasks. Extracting this structure would allow us to seamlessly reuse supervision among related tasks or solve (…)

Many practical continuous minimization problems, such as aerodynamic optimization, are not amenable to gradient-based optimization methods because derivatives cannot be computed directly. We recently showed that it is possible to train a Neural Network regressor as a proxy to the numerical simulator and optimize the proxy function via Gradient-Descent. [ICML 2018 Paper]. For example, we (…)

DescriptionUsed in a variety of computer vision tasks, mesh parameterizations are powerful tools to represent 3D shapes as a compact set of parameters (typically a vector of size 256). At CVLab, in 2020 we successfully developed such a generic mesh parameterization called MeshSDF. This project would explore ways of modernizing it, with a particular focus (…)

A network learns to predict the SDF, a volumetric function describing implicitly the surface of the shape. It takes as input a latent vector z encoding the shape and the 3D point x where the function is queried. The resulting SDF can be meshed to recover the 3D surface.DescriptionThese past years have seen an increased (…)