| 摘要 |
A method of encoding an image of a video sequence, the image being represented by a plurality of coding units, the encoding comprising at least two different encoding processes for encoding the image wherein each encoding process introduces coding dependencies between encoded data associated with said encoding process (e.g. the coding dependency of a coding unit on its neighbouring coding units in the image). The method comprises: determining a set of slice boundaries for each encoding process wherein the set of slice boundaries for each encoding process partitions the image into one or more slices for the respective encoding process. The slice boundaries for each encoding process are defined as a function of at least one of the characteristics of data content of the video sequence, the characteristics (e.g. packet loss rate and maximum transport unit size) of the network via which the encoded image is to be transmitted and the capabilities of the decoder (e.g. the number of decoding processes that can be run in parallel, error concealment capability) to which the encoded image is to be transmitted. The coding units of one slice being encoded independently of the coding units of another of the slices of the encoding process concerned. |
