Complexity-aware video coding methods for resource-constrained decoders
- Cordeiro, Paulo Jorge Ferreira Batista
- Juan Antonio Gómez Pulido Director
- Pedro A. Amado de Assunção Director
Defence university: Universidad de Extremadura
Fecha de defensa: 13 July 2015
- Óscar Gutiérrez Blanco Chair
- Vasco Nuno da Gama de Jesus Soares Secretary
- Francisco José Bellido Outeiriño Committee member
- Joel José Puga Coelho Rodrigues Committee member
- Pedro Miguel Núñez Trujillo Committee member
Type: Thesis
Abstract
This thesis addresses the problem of reducing and balancing video decoding complexity by dynamically adapting the coding process without compromising too much the coding efficiency. In this context, complexity-aware video coding has been under investigation in the last few years, particularly targeting standard video codecs like the widely used H.264/AVC (Advanced Video Coding) and its successor, the emerging HEVC (High Efficiency Video Coding). The latter achieves significant improvements in video quality, at the expense of greatly increased computational complexity at both the encoder and the decoder. The research work carried out in the scope of this thesis comprise three main scientific contributions to the generic problem mentioned above. First contribution is based on a coding mechanism which is able to generate H.264/AVC video streams with reduced decoding complexity. Second contribution addresses the most relevant challenges posed by VSN (Video Sensor Networks), namely stringent bandwidth usage and processing time/power constraints. The third contribution presents an investigation about decoding complexity of HEVC and develops a model for this standard decoder based on its fundamental coding units (CU). Overall the work done and results achieved in this thesis contribute to advance the current state of the art in the field of complexity-aware video coding, particularly focused on the most recent standard video codecs.