Jump label

Service navigation

Main navigation

You are here:

Main content

Spatio-temporal image enhancement by iterative reconstruction algorithms: artifact reduction and super-resolution

   

Research project of Circuits and Systems Lab

 

Bildinhaltsabhängige Signalverarbeitung     

The number of different digital video sources and formats strongly increased over the last years. Whereas standard definition format was the only source for TV panels some years ago, actual panels have to cope with videos featuring completely different origin, resolution, and quality. Examples are videos with QCIF or an even lower resolution typical for Web-Portals, standard definition video with different data rates (DVB-T/C/S), and high definition television or 4K video. Due to the fact that the viewer is accustomed to HDTV, enhancement of legacy content is necessary achieving at least a quality close to HD. This is a challenging task for input sources with strongly varying quality requiring highly adaptive video processing. Iterative reconstruction algorithms are a very promising processing strategy fulfilling these demands. These algorithms are based on mathematical approaches, but they can be derived into adaptive and recursive filter structures operating on pixel level, allowing a hardware efficient application capable for real-time processing. Objective of this project is to determine the potential of these algorithms, measure the achievable quality, and compare their performance to standard techniques. Additionally, potential improvements by integrating feature analysis means can further enhance the quality of these algorithms.

Bildinhaltsabhängige Signalverarbeitung 

Publications:

Richter, M.; Dolar, C.; Schröder,H.:

  • Rekursive und direkte Verfahren in der Artefaktreduktion. Angenommen zur Publikation: FKTG Jahrestagung 2010, Hamburg, Mai 2010

Richter, M.; Dolar, C.; Schröder,H.:

  • Coding Artifact Reduction by Temporal Filtering. IEEE Internat. Symposium on Consumer Electronics (ISCE 2009), Kyoto, Japan, Mai 2009

Dolar, C.; Richter, M.; Schröder,H.:

  • Total Variation Regularization Filtering for Video Signal Processing. IEEE Internat. Symposium on Consumer Electronics (ISCE 2009), Kyoto, Japan, Mai 2009

Dolar, C.; Richter, M.; Schröder,H.:

  • Total Variation Regularization for Video Signal Processing Applications. ITG Fachtagung für Elektronische Medien "Systeme, Technologien, Anwendungen", 13. Dortmunder Fernsehseminar, Dortmund, März 2009

 

Schlagworte:

Super-Resolution, Artifact Reduction, Optimization, Iterative Processing



Sub content

Contact

M.Sc. Christian Brauers