Performance Analysis of One-Bit Transform based Motion Estimation using Sparse Search on HEVC


DUVAR R., KÜÇÜKMANİSA A., AKBULUT O., TAŞYAPI ÇELEBİ A., URHAN O.

26th IEEE Signal Processing and Communications Applications Conference (SIU), İzmir, Turkey, 2 - 05 May 2018 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2018.8404776
  • City: İzmir
  • Country: Turkey
  • Kocaeli University Affiliated: Yes

Abstract

Video encoders generally employ sparse search-based approaches since the full search-based motion estimation approaches have a significant amount of computational burden. Sparse search-based approaches are often designed by taking software implementations into account performance, but hardware implementations of these methods may suffer important performance losses. In this work, the performance of the Test Zone Search approach which is the fastest motion estimation used in HEVC, is examined when it is integrated with hardware-based 1-bit transform based low bit depth motion estimation approach.