Scalable Video Coding Sample Clauses

Scalable Video Coding. The rapidly growing number of user devices with different processing capabilities and display features along with the time-varying and potentially unpredictable connection qualities raises the need for a video coding standard that provides multiple video representations with different bit rates, frame rates and frame sizes at relatively low computational cost [30–32]. In video coding, a video sequence is a series of pictures taken at a constant time intervals. Video compression techniques take advantage of the similarities between adjacent pictures to reduce the bit rate. Similar to the the well-known and widespread H.264/AVC standard [33, 34], in SVC, each video sequence is divided into one or more group of pictures (GoP), and each GoP is encoded with three different types of frames: I-frame (intra frame), P-frame (predicted frame) and B-frame (bidirectionally predicted frame). An I-frame contains an entire image and is independently coded without reference to any other frames. A P-frame is coded/decoded using the preceding P-frame or I-frame. Hence, the correct decoding of P-frames depends on the availability of the previous I- or P-frame. On the other hand, a B-frame depends on both preceding and future frames [35, 36]. Therefore, in a video sequence, I-frames constitute a greater importance in comparison with the P-frames and B-frames. SVC has been standardized as the scalable extension of H.264/AVC standard by the Joint Video Team (JVT) [37]. In contrast to single-layer bit streams, in a scalable video stream, a lower quality/resolution version of the video can be cre- ated just by dropping packets and removing parts of the complete original stream without re-encoding [2, 38, 39]. SVC’s capability in reconstructing lower resolu- tion or lower quality signals from bit streams without the need for re-encoding allows for simple solutions in adaptation to network and terminal capabilities [40]. However, adaptation operations, which consist of an adaptation decision and a thinning operation to discard unneeded data are very complex operations [41]. SVC incorporates three scalability modes, providing a full scalability in tem- poral, quality and spatial dimensions [2, 39, 41, 42]. Temporal scalability allows dividing the video stream frames into a base layer which is coded to provide the basic frame rate and one or multiple enhancement layers. Enhancement layers are coded using temporal prediction corresponding to the base layer or higher enhancement layers...