Management of slices

摘要

A slice (3) of a picture (1) is encoded by hierarchically splitting a largest coding unit, LCU, (10) of the picture (1) into multiple smaller coding units, CUs (20, 30, 40). A start of the slice (3) is defined to coincide with the border between a first CU (30, 40B) and a preceding CU (40, 40A) of the LCU (10). Address information is generated that allows identification of the first CU (30, 40B) within the picture (1) by defining the position of the first CU (30, 40B) and the size of the largest possible CU (30) that can occupy this position within the picture (1). The address information is included in a coded representation (50) of the picture (1). The address information enables positioning slice starts also within LCUs (10), which slice starts thereby do not need to be limited between LCU borders.

主权项

1. A method of coding a picture comprising multiple slices and being composed of multiple largest coding units having a selected size in terms of number of pixels, said method comprising:
hierarchically splitting a largest coding unit of said multiple largest coding units into multiple coding units having respective sizes that are smaller than said selected size; defining a start of a slice in said picture to coincide with a border between a first coding unit of said multiple coding units and a preceding, according to a defined processing order, coding unit of said multiple coding units; selecting the first coding unit and the preceding coding unit to have different sizes in terms of number of pixels. wherein the first coding unit and the preceding coding unit belong to a same largest coding unit: generating address information allowing identification of said first coding unit, said address information defining a position of said first coding unit within said picture and a size of a largest possible coding unit that can occupy said position in said picture; and including said address information in a coded representation of said picture.