The goal of this project is to develop cost-effective methods for camera pose estimation. The proposed methods are useful in areas such as autonomous driving, UAVs, augmented reality.

The goal of this project is to improve current state of the art 3d scanning methods to increase robustness and reduce price.

The goal of this project is to propose new methods for Augmented Reality applications.

The goal of this project is to detect and estimate the 2D pose of humans in stereo image pairs from realistic stereo videos.

The goal of this project is to estimate the 3D location of human body parts in a multicamera setup without using markers.

The goal of this project is to automatically recognize person-person interactions. We address this problem by combining visual and audio sources.

The goal of this project is to recognize individuals by analysing their gait.

The goal of this project is to design a novel method for procedurally modelling large, complex three–dimensional scenes. Our approach is general-purpose and takes as input any three–dimensional model intuitively provided by a user. The method exploits the adjacency between shapes and objects in the input model and computes an output model that extracts these features (constraints and adjacencies) and models the input. We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Geforce GTX Titan X GPU used for this research.

The goal of this project is to develop efficient and robust people detection and tracking methods using depth sensors.