Videocodierung

Topics in the area of combining dictionary learning / sparse coding with video coding Jens Schneider M.Sc.

The field of dictionary learning and sparse coding has raised more and more interest in the last years. State of the art results for various problems of signal processing were obtained by these methods. More precisely, dictionary learning shows good performance for denoising, inpainting and super-resolution. Therefore, it is assumed that dictionary learning / sparse coding methods can also be used for compression of video content. In possible thesis work it will be investigated how sparse coding methods can be introduced into the well known hybrid video coding scheme consisting of intra prediction, inter prediction and residual coding.

Possible topics in this area lie in the field of:

  • Dictionary learning / sparse coding for intra prediction in video coding
  • Dictionary learning / sparse coding for dynamic resolution conversion coding
  • Optimization of sparse coding algorithms

Every thesis will require implementations in MATLAB or C++. Therefore, a high motivation and interest for software programming is recommended.

Masterthesises in the area of neural network based video intra prediction Maria Meyer M.Sc.

A fundamental part of each image or video coding algorithm are methods that try to predict parts of the frames or images based on the information that was coded beforehand in order to significantly reduce the redundancy. These methods can be classified into two basic categories: Inter prediction methods that use information from previous frames and intra prediction methods that only information from the local neighborhood within the same frame.
While conventional intra prediction methods determine the prediction based on a set of linear filters, some recent proposals started to use neural networks for this purpose. In particular it is possible to train a convolutional neural network (CNNs) to predict a block based on its local neighborhood and use the result as an additional prediction option in a video coder. Unlike the conventional methods, these CNNs can usually adopt to nonlinear edges and textures. However, there is still much room for improvement in the training techniques and inference optimization processes.
There are frequently master thesis topics coming up in this general area. Please ask, if you are interested in the general area.