This book focuses on the research and development challenges posed by 3D video systems based on multi-view plus depth (MVD) technology. This technology can produce a realistic immersive experience generating synthetic video views on the decoder side, reducing the amount of information on the encoder side. The discussion presented in this book explores the MVD characteristics to propose high-throughput and energy-efficient architectures/systems, focusing on 3D-HEVC, the state-of-the-art standard for exploiting the MVD concept. The book includes an extensive discussion of the 3D-HEVC video encoding, followed by an in-depth evaluation of the 3D-HEVC reference software behavior. Then, the book presents in detail a set of high-throughput and energy-efficient architectures targeting the three main prediction steps inside the 3D-HEVC: intra-frame prediction, inter-frame prediction, and inter-view prediction.
In addition, this book includes:
- Provides a detailed description of the new MVD encoding tools used in 3D-HEVC.
- Offers an overview of state-of-the-art works on MVD hardware architectures.
- Includes high-throughput architectures for inter-prediction focused on Inter-frames and Inter-view predictions.
