KR20080008205A - Method for encoding and decoding video signal - Google Patents

Abstract

A method for encoding/decoding a video signal is provided to improve coding efficiency in interlayer prediction by using motion information included in an FGS(Fine Granular Scalability) layer of a base layer. A video signal encoding method comprises the following steps. By using motion information of a lower layer encoded form the same video signal, a bit stream of a current layer is generated. The lower layer includes a first layer having one or more levels for compensating loss generated during encoding of a lower level. A predetermined level including the motion information in the first layer is used for generating the bit stream of the current layer.

Description

Method for encoding and decoding video signal

1 conceptually illustrates a conventional inter-layer motion prediction method using only a quality base of a base layer,

FIG. 2 conceptually illustrates an embodiment of the present invention for predicting motion for quality base of a current layer using an FGS layer of a base layer,

3 conceptually illustrates another embodiment of the present invention for predicting motion for an FGS layer of a current layer using an FGS layer of a base layer,

4 illustrates a configuration of an encoding apparatus to which an encoding method of a video signal according to the present invention is applied.

FIG. 5 shows a configuration of an apparatus for decoding a bit stream encoded by the apparatus of FIG. 4.

<Explanation of symbols for the main parts of the drawings>

110: BL encoder 120: EL encoder

130: Eat 210: Demux

220: BL decoder 230: EL decoder

The present invention relates to a method for encoding and decoding a video signal.

The scalable video codec (SVC) method encodes a video signal, and encodes at the highest quality, but decodes a partial sequence of the resultant picture sequence (a sequence of intermittently selected frames in the entire sequence). It is a method to allow a certain level of image expression even if used.

A picture sequence encoded in a scalable manner can be represented to a certain degree of image quality by receiving and processing only a partial sequence. However, when a bit rate is lowered, image quality deterioration is large. In order to solve this, a separate auxiliary picture sequence for a low data rate, for example, a small picture and / or a low picture sequence per frame may be provided.

The auxiliary picture sequence is called a base layer, and the main picture sequence is called an enhanced layer or an enhancement layer. The base layer and the enhanced layer encode the same video signal source, and redundant information (redundancy) exists in the video signals of the two layers. Therefore, when a base layer is provided, an inter-layer prediction method for predicting and encoding an image signal of an enhanced layer using motion information and / or texture information, which is image data, of the base layer to improve coding efficiency. (interlayer prediction method) can be used.

In addition, an enhanced layer may be used to improve the signal-to-noise ratio (SNR) of the base layer, that is, to improve image quality, which may be used for SNR scalability, fine granular scalability (GFS), or incremental refinement (FGS). Progressive Refinement).

According to the FGS, transform coefficients corresponding to each pixel, for example, DCT (Discrete Cosine Transform) coefficients, are divided into a base layer (also called a quality base) and an FGS layer according to the resolution of the bit representation. If the transmission environment is bad, transmission of the FGS layer may be omitted, thereby lowering the bit rate while reducing the quality of the decoded video. That is, the FGS layer compensates for the loss occurring during the quantization of the quality base, and provides a high flexibility of controlling the bit rate in response to a transmission or decoding environment.

For example, if the base layer is generated by quantizing the transform coefficients to a quantization step size (aka QP), for example QP = 32, then FGS layer level 1 (aka FGS level 1 (FGS L1)) is the original transform coefficient. The difference between the transform coefficient obtained by inverse quantization of the quantized coefficient of the base layer (quality base) and the quantization step size corresponding to a higher quality than QP 32 is generated by quantization, for example, QP = 26. Similarly, FGS level 2 (FGS L2) quantizes, for example, QP = 20 for the difference between the original transform coefficient and the transform coefficient obtained by inverse quantization of the sum of the quality base and the quantized coefficient of FGS level 1. Is generated.

Meanwhile, inter-layer motion prediction during inter-layer prediction predicts and encodes an image signal of an enhanced layer, that is, a current layer, using motion information of the base layer.

In addition, the current Joint Scalable Video Model (JSVM) allows motion information to be transmitted to FGS layers such as FGS L1 and FGS L2 so that redundant information can be efficiently used in connection with FGS. Motion information transmitted to the FGS layer, which provides higher image quality compared to quality base, is more accurate than motion information based on quality.

However, as shown in FIG. 1, the current inter-layer motion prediction does not use the motion information of the FGS layer and has low precision even when there is one or more FGS layers (levels) including the motion information in the base layer. Only the base motion information is used.

The present invention was created in view of the above circumstances, and an object of the present invention is to improve coding efficiency in inter-layer prediction, and to provide a method of using motion information included in an FGS layer of a base layer.

In order to achieve the above object, a method of encoding a video signal according to an embodiment of the present invention generates a bit stream of a current layer by using motion information of a lower layer encoded from the same video signal. A first layer having one or more levels for compensating for losses occurring in lower level encoding, wherein a predetermined level comprising motion information in the first layer is used to generate the bit stream of the current layer It is done.

In addition, the method of decoding an encoded video bit stream according to another embodiment of the present invention, decoding the bit stream of the lower layer for the same video signal, using the motion information of the decoded lower layer bits of the current layer Decode a stream, wherein the lower layer comprises a first layer having one or more levels for compensating for losses incurred in lower level encoding, wherein the predetermined layer containing motion information in the first layer is the current layer; It is characterized in that it is used for decoding.

In an example of the present invention, information indicating the lower layer and the level in the lower layer is recorded in the header area of the current layer, and the level indicated by the information is lower than the level indicated by the information if the level does not include motion information. A lower level with motion information may be used to generate the bit stream of the current layer. The motion information includes a motion vector.

In addition, according to another embodiment of the present invention, a method of encoding a video signal generates a bit stream of a current layer by using a lower layer encoded from the same video signal, wherein the current layer and the lower layer each have a lower level. First and second layers having one or more levels to compensate for losses arising in the encoding, wherein any level of the first layer is a lower level of motion information in the first layer or a predetermined level in the second layer. It is encoded using the motion information of the level of, characterized in that the information indicating which motion information of the current layer or the lower layer is used for encoding any level of the first layer is characterized in that it is recorded.

In addition, the method of decoding an encoded video bit stream according to another embodiment of the present invention, decoding the bit stream of the lower layer for the same video signal, and decodes the bit stream of the current layer using the decoded lower layer Wherein the current layer and the lower layer each comprise a first and a second layer having one or more levels to compensate for the loss arising in the lower level of encoding, and are present in decoding any level of the first layer. Information indicating which motion information of a layer or a lower layer is used is read out, and based on this, any level of the first layer is based on motion information of a lower level in the first layer or a predetermined level in the second layer. It is characterized in that the decoding using the level of the motion information.

In one example of the present invention, the information may be defined in a slice unit corresponding to a level or in an image block unit in a level. In addition, when the information indicates that motion information of a lower layer is used, a predetermined level used for encoding the arbitrary level is selected from a lower layer and a level used for generating a bit stream of the current layer except for the first layer. It is derived that if the level n in the second layer of the lower layer used to generate the bit stream of the current layer except for the first layer does not include motion information, the lower level lower than the level n and having the motion information is the arbitrary level. Can be used for encoding of levels.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

According to the current JSVM, in inter-layer texture prediction during inter-layer prediction, the video signal of the current layer is encoded not only by the quality base of the base layer but also by using the FGS layer of the base layer. Information including the FGS layer of the base layer, for example, base_id_plus1 is included in the slice header and transmitted.

The decoder determines the base layer to be used for inter-layer prediction from base_id_plus1 included in the slice header of the bit stream, and also provides information for identifying a quality base or FGS layer (quality level) in the determined base layer, for example, QualityBaseLevel. A quality base or an FGS layer to be used for inter-layer prediction is determined from the base_id_plus1. However, this information has not been used for inter-layer motion prediction.

As information about the base layer and the quality level are transmitted for the inter-layer texture prediction as described above, in the present invention, the FGS layer of the base layer as well as the quality base of the base layer can be used by using such information without additional transmission of new information. Inter-layer motion prediction may be performed based on the motion information.

In an embodiment of the present invention, as shown in FIG. 2, base_id_plus1 defined in a slice header indicates whether a motion vector of which FGS level of which layer is used for quality based motion prediction of a current layer (enhanced layer). Check based on QualityLevelBase included in base_id_plus1.

The base_id_plus1 is defined in units of slices, and includes predetermined bits indicating which base layer is used for inter-layer prediction of macroblocks in the slice, for example, information of 4 bits, and indicating which FGS layer is used in the base layer. Bits, for example, two bits of information (QualityBaseLevel) and the like.

The base layer below the current layer (enhanced layer) may further include 1 to 3 FGS layers FGS L1 to L3, respectively, and FGS L0 corresponds to the quality base of the base layer.

Therefore, if QualityLevelBase included in base_id_plus1 in the slice header is 0, motion and / or texture information of the quality base FGS L0 of the base layer indicated by base_id_plus1 is used for inter-layer prediction for the current layer. Similarly, if the QualityLevelBase is 11, motion and / or texture information of the quality base (FGS L3) of the base layer indicated by the base_id_plus1 is used for inter-layer prediction for the current layer.

In this case, when motion information is not included in the FGS layer of the base layer indicated by the base_id_plus1 and QualityLevelBase included therein, the motion information of the quality base of the base layer indicated by the base_id_plus1 may be used for inter-layer motion prediction for the current layer. Alternatively, motion information of another FGS layer including motion information below the FGS layer indicated by the QualityLevelBase may be used for inter-layer motion prediction for the current layer.

In another embodiment of the present invention, as shown in FIG. 3, the motion information for the FGS layer of the current layer (enhanced layer) is predicted using the motion information of the FGS layer of the base layer.

In the current JSVM, texture and / or motion vectors are predicted using only the lower level FGS layer (or quality base) of that layer (the current layer) for the FGS layer, so prediction with other layers, i.e. inter-layer prediction In this case, base_id_plus1 information is not transmitted to the slice of the FGS layer.

Therefore, for inter-layer motion prediction for the FGS layer of the current layer, the motion information of the FGS layer (or quality base) of the lower level of the layer (current layer) or the FGS layer of another layer (base layer) is used. Information indicating whether or not, for example, motion_base_prediction_flag of 1 bit should be transmitted. The motion_base_prediction_flag may be defined in macroblock units or slice units.

The motion_base_prediction_flag is given a value of '0', for example, when the motion of the FGS layer of the current layer is predicted using the motion information of the FGS layer (or the quality base) of the lower level of the current layer. When the motion of the FGS layer of the current layer is predicted using the motion information of the FGS layer of the base layer), for example, a value of '1' may be given.

When the motion information of the FGS layer of another layer (base layer) is used, that is, motion_base_prediction_flag is '1', the quality of the current layer to determine which base layer and which FGS layer are used to predict the motion of the FGS layer of the current layer. It may be determined based on base_id_plus1 information defined in the base slice.

4 illustrates a configuration of an encoding apparatus to which an encoding method of a video signal according to the present invention is applied.

The encoding apparatus includes an enhanced layer encoder (EL Encoder) 120 for generating an enhanced layer for an image signal through inter-layer prediction based on a data stream of a base layer, and a data stream of the generated enhanced layer; And / or a muxer 130 for encapsulating the data stream of the base layer in a predetermined format and then muxing and outputting each other in a predetermined transmission format.

A data stream of a base layer already generated for another use by another encoding device may be provided to the encoding device for generation of an enhanced layer through inter-layer prediction. Alternatively, the encoding apparatus may further include a base layer encoder 110 generating a base layer of an image signal of the same source, and may directly generate a base layer to use for inter-layer prediction.

The muxer 130 muxes and outputs the base layer and the enhanced layer, considering that the decoding apparatus can obtain the data stream of the base layer through a path separate from the data stream of the enhanced layer. Only data streams of the enhanced layer generated by the EL encoder 120 may be muxed.

When encoding the quality base of the current layer, the EL encoder 120 uses the FGS layer as well as the quality base of the lower layer (base layer) to perform inter-layer prediction, that is, inter-layer texture prediction and / or inter-layer motion prediction. The base layer used for inter-layer prediction and information (base_id_plus1) indicating the FGS layer (or the quality base) of the base layer are inserted into the quality base slice header of the current layer.

The EL encoder 120 uses any FGS level (layer) of any base layer in inter-layer texture prediction for the quality base of the current layer, but if the any FGS level does not include motion information, In inter-layer motion prediction for a quality base of a layer, motion information of an FGS level including motion information of the quality base of the arbitrary base layer or motion information below the arbitrary FGS level may be used. In this case, the EL encoder 120 inserts information (base_id_plus1) indicating the base layer and the FGS level used in the inter-layer texture prediction in the quality base slice header of the current layer.

In addition, the EL encoder 120 uses motion information of an FGS layer (or quality base) below the current layer or encodes another layer (base layer) when encoding while providing motion information to the FGS layer of the current layer. Motion prediction between layers can be performed using the motion information of the FGS layer (or quality base) of.) A macroblock header or FGS is defined by defining one bit information (motion_base_prediction_flag) in units of macroblocks or slices. Insert into the layer slice header.

The encoded data stream is wired or wirelessly transmitted to the decoding apparatus or transmitted via a recording medium, and the decoding apparatus restores the original video signal according to the method described later.

FIG. 5 illustrates a configuration of an apparatus for decoding a bit stream encoded by the apparatus of FIG. 4.

The decoding apparatus includes a demuxer 210 that separates and outputs an enhanced layer and a data stream of a base layer from a bit stream received in a transmission format, and restores an image signal of the base layer from a data stream of the base layer. The base layer decoder (BL Decoder) 220 and the base layer to restore the image signal of the enhanced layer from the data stream of the enhanced layer output from the demuxer 210, the base layer having a different resolution, frame rate, etc. And an enhanced layer decoder (EL Decoder) 230 which reconstructs an image signal of the enhanced layer through inter-layer prediction based on texture and / or motion information.

The data stream of the base layer may be provided muxed with the data stream of the enhanced layer or may be provided through a separate path. For example, the base layer may be provided via a storage medium such as DVD, Blu-ray, and the enhanced layer associated with the base layer may be provided via a wired or wireless network or through another storage medium.

When the EL decoder 230 decodes the quality base of the enhanced layer (current layer) encoded through inter-layer motion prediction, the lower layer (base) used for inter-layer prediction in the quality base slice header of the current layer. Layer) and information (base_id_plus1) indicating the FGS layer (level) of the base layer are read first, and the base layer and the corresponding FGS level (QualityLevelBase) are checked therefrom, and the motion information of the FGS level of the checked base layer is used. Inter-layer motion prediction is performed on the quality base of the current layer.

When the QualityLevelBase derived from the base_id_plus1 is 0, for example, the EL decoder 230 performs inter-layer motion prediction using motion information of the quality base of the base layer indicated by the base_id_plus1, and the QualityLevelBase is, for example. If 3, inter-layer motion prediction is performed using motion information of FGS level 3 (FGS L3) of the base layer indicated by base_id_plus1.

In addition, the EL decoder 230 uses the quality information of the base layer of the base layer, or the FGS level indicated by the QualityLevelBase, when the FGS level of the base layer indicated by the base_id_plus1 and the QualityLevelBase does not include the motion information. Inter-layer motion prediction may be performed on the quality base of the current layer by using motion information of the FGS level including motion information below.

Meanwhile, when decoding the FGS layer of the current layer, the EL decoder 230 checks motion_base_prediction_flag information in the FGS layer slice header of the current layer, and when the information is 0, the FGS layer (or the lower level of the current layer) Motion information of the current layer is used to predict the motion of the FGS layer of the current layer, whereas when the information is 1, motion information of the FGS layer of the other layer (base layer) is used to determine the motion of the FGS layer of the current layer. Predict motion. The base layer and the FGS layer to be used when the information is 1 may be determined based on base_id_plus1 information defined in the quality base slice of the current layer.

Alternatively, the motion_base_prediction_flag information may be defined in macroblock units, and when the EL decoder 230 decodes any macroblock in the FGS layer of the current layer, checks the motion_base_prediction_flag information in the header of the arbitrary macroblock. According to the value, using the motion information of the block (corresponding block) corresponding to the arbitrary macroblock in the quality base of the current layer or the motion information of the corresponding block in the FGS layer of the base layer for the arbitrary macroblock You can predict motion.

The EL decoder 230 predicts the quality base of the current layer and / or the motion of the FGS layer based on the motion information of the FGS layer of the base layer through the above operation, and uses the quality base of the current layer and / or Alternatively, the video signal of the FGS layer is restored.

The decoding apparatus according to the present invention may be mounted in a dedicated playback terminal or a terminal to which a communication function has been given, or in a device for reproducing a recording medium.

As described above, preferred embodiments of the present invention have been disclosed for the purpose of illustration, and those skilled in the art can improve, change, and further various embodiments within the technical spirit and the technical scope of the present invention disclosed in the appended claims. Replacement or addition may be possible.

Therefore, a more flexible inter-layer motion prediction method can be provided, thereby improving coding efficiency.

Claims (16)

A bit stream of the current layer is generated using motion information of a lower layer encoded from the same video signal. The lower layer includes a first layer having one or more levels for compensating for losses incurred in lower level encoding, wherein a predetermined level comprising motion information in the first layer is used to generate the bit stream of the current layer. A method for encoding a video signal, characterized in that it is used. The method of claim 1, And the information indicating the lower layer and the level in the lower layer is recorded in the header area of the current layer. The method of claim 2, And when the level indicated by the information does not include motion information, a lower level lower than the level indicated by the information and having motion information is used to generate a bit stream of the current layer. The method of claim 1, And wherein the motion information comprises a motion vector. A bit stream of the current layer is generated by using a lower layer encoded from the same video signal. The current layer and the lower layer each include a first layer and a second layer having one or more levels for compensating for losses occurring in the encoding of the lower level, Any level of the first layer is encoded using motion information of a lower level in the first layer or motion information of a predetermined level in the second layer, And information indicating which layer of the current layer or the lower layer is used for encoding any level of the first layer. The method of claim 5, And the information is defined in a slice unit corresponding to a level or in an image block unit within a level. The method of claim 6, When the information indicates that motion information of a lower layer is used, a predetermined level used for encoding the arbitrary level is derived from a lower layer and a level used for generating a bit stream of the current layer except for the first layer. A method of encoding a video signal, characterized in that. The method of claim 7, wherein When level n in the second layer of the lower layer used to generate the bit stream of the current layer except for the first layer does not include motion information, a lower level lower than the level n and having motion information may be A method of encoding a video signal, characterized by being used for encoding. The method of claim 5, And wherein the motion information comprises a motion vector. Decode the bit stream of the lower layer for the same video signal, Decode the bit stream of the current layer by using the motion information of the decoded lower layer, The lower layer includes a first layer having one or more levels for compensating for losses incurred in lower level encoding, wherein a predetermined level comprising motion information in the first layer is used for decoding the current layer. A method of decoding an encoded video bit stream. The method of claim 10, And information indicating the lower layer and the level in the lower layer is read from the header area of the current layer. The method of claim 11, And when the level indicated by the information does not include motion information, a lower level lower than the level indicated by the information and having motion information is used for decoding the current layer. Decode the bit stream of the lower layer for the same video signal, Decode the bit stream of the current layer by using the decoded lower layer, The current layer and the lower layer each include a first layer and a second layer having one or more levels for compensating for losses occurring in the encoding of the lower level, Information indicating which of the current layer or the lower layer motion information is used for decoding any level of the first layer is read, and based on this, any level of the first layer is a lower level in the first layer. A method of decoding an encoded video bit stream, characterized in that it is decoded using the motion information of a level or the motion information of a predetermined level in the second layer. The method of claim 13, Wherein the information is defined in units of slices corresponding to levels or units of image blocks within the levels. The method of claim 14, When the information indicates that motion information of a lower layer is used, a predetermined level used for decoding the arbitrary level is derived from a lower layer and a level used for generating a bit stream of the current layer except for the first layer. A method of decoding an encoded video bit stream. The method of claim 15, When level n in the second layer of the lower layer used to generate the bit stream of the current layer except for the first layer does not include motion information, a lower level lower than the level n and having motion information may be A method for decoding an encoded video bit stream, characterized in that it is used for decoding.

Images