Image decoding method using information on a random access picture and apparatus using same
The present invention relates to an image decoding method and to an apparatus using same. The image decoding method includes the steps of: receiving a bitstream including random access image information; and predicting the random access image and an image following the random access image in terms of a decoding sequence on the basis of the random access image information.
1. A video decoding method comprising:
receiving, by a decoding apparatus, a bitstream comprising information on a random access picture in a current layer;
predicting, by the decoding apparatus, the random access picture and a subsequent picture following the random access picture in decoding order in the current layer on the basis of the information on the random access picture,
wherein the predicting comprises deriving a reference picture for the random access picture and the subsequent picture on the basis of at least one of a temporal layer identifier to identify a temporal layer, a spatial layer identifier to identify a spatial layer, and a quality layer identifier to identify a quality layer,
wherein in the predicting, when inter prediction is performed for the random access picture and the subsequent picture, a picture having a temporal layer identifier value the same as or greater than a temporal layer identifier value of the random access picture is not used as a reference picture; and
generating, by the decoding apparatus, reconstructed pictures based on the result of the predicting,
wherein the information on the random access picture comprises information specifying whether a current picture is a cross-layer random access point at which switching from a first layer to a second layer is possible,
wherein the first layer and the second layer are a first spatial layer and a second spatial layer for supporting spatial scalability, or a first quality layer and a second quality layer for supporting quality scalability, wherein the first layer is a lower layer than the second layer,
wherein in the predicting,
when a corresponding picture, of the random access picture, in the first layer is a random access point picture, inter-layer prediction is available for the random access picture in the current layer which is the second layer, and
when the corresponding picture in the first layer is not a random access point picture, inter-layer prediction is not available for the random access picture in the current layer which is the second layer.
2. The video decoding method of claim 1 , wherein in the predicting, when inter-layer prediction is performed for the random access picture and the subsequent picture, a picture having a spatial layer identifier value smaller than or the same as a spatial layer identifier value of the random access picture or a picture having a quality layer identifier value smaller than or the same a quality layer identifier value of the random access picture is used as a reference picture.
3. The video decoding method of claim 1 , wherein the information on the random access picture further comprises information on a random access picture type, and the information on the random access picture type is defined as a network abstraction layer (NAL) unit type.
4. The video decoding method of claim 1 , wherein the information on the random access picture is stored as syntax information.
5. The video decoding method of claim 1 , wherein the random access picture is one of a spatial layer access or scalable layer access (SLA) picture, a layer access (LA) picture and a clean random access (CRA) picture.
6. A video decoding apparatus comprising:
an entrophy decoder to receive a bitstream comprising information on a random access picture in a current layer;
a predictor to predict the random access picture and a subsequent picture following the random access picture in decoding order in the current layer on the basis of the information on the random access picture,
wherein the predicting comprises deriving a reference picture for the random access picture and the subsequent picture on the basis of at least one of a temporal layer identifier to identify a temporal layer, a spatial layer identifier to identify a spatial layer, and a quality layer identifier to identify a quality layer,
wherein in the predicting, when inter prediction is performed for the random access picture and the subsequent picture, a picture having a temporal layer identifier value the same as or greater than a temporal layer identifier value of the random access picture is not used as a reference picture; and
an adder to generate reconstructed pictures based on the result of the prediction,
wherein the information on the random access picture comprises information specifying whether a current picture is a picture is a cross-layer random access point at which switching from a first layer to a second layer is possible,
wherein the first layer and the second layer are a first spatial layer and a second spatial layer for supporting spatial scalability, or a first quality layer and a second quality layer for supporting quality scalability, wherein the first layer is a lower layer than the second layer,
wherein in the predicting,
when a corresponding picture, of the random access picture, in the first layer is a random access point picture, inter-layer prediction is available for the random access picture in the current layer which is the second layer, and
when the corresponding picture in the first layer is not a random access point picture, inter-layer prediction is not available for the random access picture in the current layer which is the second layer.