Artificial Intelligence

Blog Post
Towards Face-to-Face Telepresence Applications in Virtual Reality with Generative Neural Nets
Philipp Ladwig held a talk at the 17th GI VR/AR workshop about our research in the area of telepresence in virtual reality applications. The associated paper written by Philipp Ladwig and Alexander Pech under the guidance of Prof. Dr. Christian Geiger received the "Best Paper" award.
Publication
Smart Object Segmentation to Enhance the Creation of Interactive Environments
The objective of our research is to enhance the creation of interactive environments such as in VR applications. An interactive environment can be produced from a point cloud that is acquired by a 3D scanning process of a certain scenery. The segmentation is needed to extract objects in that point cloud to, e.g., apply certain physical properties to them in a further step. It takes a lot of effort to do this manually as single objects have to be extracted and post-processed. Thus, our research aim is the real-world, cross-domain, automatic, semantic segmentation without the estimation of specific object classes.
Publication
Towards Face-to-Face Telepresence Applications in Virtual Reality with Generative Neural Nets
Three-dimensional capture of faces with personal expressions for facial reconstruction under a head-mounted display.
Publication
Point Cloud Segmentation: Solving a Perceptual Grouping Task with Deep Reinforcement Learning
We propose a method to segment a real world point cloud as perceptual grouping task (PGT) by a deep reinforcement learning (DRL) agent. A point cloud is divided into groups of points, named superpoints, for the PGT. These superpoints should be grouped to objects by a deep neural network policy that is optimised by a DRL algorithm.
Publication
Point Cloud Segmentation with Deep Reinforcement Learning
The segmentation of point clouds is conducted with the help of deep reinforcement learning (DRL) in this contribution. We want to create interactive virtual reality (VR) environments from point cloud scans as fast as possible. These VR environments are used for secure and immersive trainings of serious real life applications such as the extinguishing of a fire. It is necessary to segment the point cloud scans to create interactions in the VR
Blog Post
Personalized Avatars Generated by a Neural Network for Telepresence and Mixed Reality
Skype, Facetime and similar apps have become an integral part of our everyday life. However, the digital meeting does not compare to the real, physical one. Both spatial perception and non-verbal communication are limited by the current digital channels.
Blog Post
Photogrammetry
Our goal is to create realistic 3D images of objects or rooms. For this purpose, we use the laser scanner Faro Focus S70, which can capture anything visible in a sphere with a radius of up to 70 meters. Whenever we scan a room, we also move along the walls and take overlapping photographs. Overall, depending on the size of the room or object, we collect between 600 and 1500 pictures to ensure that we obtain enough material in a single scanning process.
Publication
Fingertracking durch neuronale Netze anhand reduzierter Markersets und Motion-Capture-Daten
In diesem Paper wird ein neuartiger Ansatz eines Fingertracking-Systems auf Basis neuronaler Netze vorgestellt. Dazu werden mithilfe eines externen, optischen Trackingsystems kleine Marker (Ø 4 mm) an den Fingerspitzen erkannt, womit zwei neuronale Netze die Positionen aller Fingergelenke bestimmen.

No posts could be found.