Navigation

Verzerrungskorrigierte Fisheye Videokomprimierung mit VVC

Verzerrungskorrigierte Fisheye Videokomprimierung mit VVC

Beschreibung

Video compression algorithms are critical for efficiently storing and transmitting the
vast amounts of data that go along with high quality video signals. An uncompressed
4K video signal at 30 frames per second leads to a datarate of roughly 712 MB/s for
conventional 8bit RGB color precision. For a 2 hour movie, this would lead to more
than 5 TB of data. State-of-the-art video compression algorithms are able to reduce
the amount of data considerably and are key to todays’ multimedia experience.

Fisheye lenses provide major benefits for many applications due to their large field
of view. However, they come at the cost of strong radial distortions leading to
problems in a variety of signal processing tasks which have been developed with
perspective lenses in mind. As such, while the state-of-the-art H.265/HEVC video
coding standard excels in reducing redundancy and irrelevance in content captured
with perspective lenses, the coding gain reduces significantly when fisheye lenses are
applied [1].

In this thesis, the effect of the distortions in fisheye video sequences on the coding
performance of the upcoming Versatile Video Coding (VVC) standard shall be eval-
uated. You will develop and implement a testing framework that allows to undistort
the fisheye video sequences either before encoding or after decoding the data. You
will then compare the coding performance of both approaches – undistortion before
encoding and undistortion after decoding – and perform a careful and meaningful
evaluation of the results stating the effect that the radial distortions have on the
upcoming VVC standard. The thesis can be written in German or English.

Distortion-corrected Fisheye VVC Test Concept

Referenzen

  1. A. Eichenseer, and A. Kaup, „Coding of Distortion-Corrected Fisheye Video Sequences using H.265/HEVC,“ in IEEE Int. Conf. Image Process., Oct 2014, pp. 4132–4136

Voraussetzungen

  • experience in programming languages such as Python
  • strong motivation for learning and mastering new tools

Betreuer

Andy Regensky, M.Sc.
andy.regensky@fau.de

Raum 06.0192
Cauerstr. 7
91058 Erlangen

Hochschullehrer

Prof. Dr.-Ing. André Kaup
andre.kaup@fau.de

Raum 06.031
Cauerstr. 7
91058 Erlangen