Projects
Video Processing for Multiview Multimodal Camera Systems Ghent University
Free viewpoint (multiview) video aims at enabling virtual navigation in a dynamic 3D scene from
arbitrary camera locations and orientations. The associated technologies also signify the solution to
the problem of providing genuinely immersive “lasses-free”3D media, offering not only stereopsis,
but also motion parallax. Capturing high-quality multiview content from real-world scenes is an
enormous ...
research on a scalable architecture for video processing embedded in an IP-network Ghent University
Teledistribution companies are increasingly interested in "thin client" concepts, in which video processing is embedded in the IP-network. This requires new, scalable and high-performance computing, which could be found in "massively multi-core processing". As hardware, algorithms and instructions can no longer be optimized separately, the problem is very complex. A single-core architecture will be developed, which could be part of a scalable ...
ASPRO+: AutoStereoscopic multi-view video PROcessing and PROjection Hasselt University
High Level Synthesis Framework for Fast Prototyping of Embedded Video Processing Pipelines Hasselt University
Deep-Learning based Joint Audio, Video Processing for Augmented Listening Ghent University
Augmented listening implies the extraction of desired audio signal(s) from a distorted capture. Like human perception of speech, where visual and acoustic cues jointly contribute to the understanding, we aim to improve this extraction by augmenting the audio with visual information. A side-application is the detection of inconsistent streams, hinting at deepfakes or otherwise compromised streams.
Compression History Discovery through Video Range Decoding for Fake Video Detection Ghent University
Imagine a video of Zelenskyy on the internet in which he makes false statements. This would have devastating consequences, yet it could be fake. Is it really the President, or is it a fake? This proposal aims to answer this question. To detect manipulation, we exploit “compression history fingerprints”. This concept considers that fake regions typically underwent a different amount of compression steps, e.g., if a fake face was added to a ...
Optimization of automatic video surveillance by multimodal video analysis Ghent University
By merging the different kinds of data provided by visual, thermal and/or depth cameras, automatic video surveillance can achieve more accurate object detection. Since, besides the visual information, also thermal and depth information can be extracted, object identification can probably also be improved. New multimodal object detection and recognition techniques will be investigated to achieve these goals.
VIDA: Video and Activate - The use of video in a process of cross-media production and activation AP Hogeschool Antwerpen
Large Scale Video Understanding KU Leuven
In this dissertation, we address the challenge of video understanding and propose several promising methods for its solution. In the context of video analysis, the recognition of human actions, the detection of objects present, and the classification of the observed scene are all important tasks. Good application examples for such semantic video analysis are more effective video retrieval, summarization, or recommendation tools. With those ...