KR101253156B1 - Method for encoding/decoding video signal - Google Patents

Abstract

The present invention relates to a method for encoding and decoding a video signal. In one embodiment according to the present invention, a deblocking filter is always applied to an image frame of a base layer to be used as a reference image for prediction of an enhanced layer. Further, in another embodiment according to the present invention, the intensity of the deblocking filter to be applied to the image after the decoding process is adjusted based on the spatial resolution of the base layer and the enhanced layer, and the ratio of the resolution of the enhanced layer and the base layer is If not 2, a high intensity deblocking filter is applied. Therefore, it is possible to make up for insufficient and incorrectly defined parts of the current JSVM deblocking filter process.

Deblocking filter, intra base mode, filter strength, dyadic, base layer, enhanced layer

Description

Method for encoding / decoding video signal {Method for encoding / decoding video signal}

1 is a schematic representation of the encoding algorithm of AVC,

2 illustrates an example in which a deblocking filter is applied to vertical or horizontal edges of an image block.

3 illustrates a process of selecting a boundary intensity parameter when applying a deblocking filter to an image after a decoding process according to an embodiment of the present invention.

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.

SVC is a codec extended from AVC (Advanced Video Codec: also referred to as 'H.264'), and FIG. 1 schematically illustrates an encoding algorithm of AVC. AVC decoding proceeds in the reverse order of the encoding process of FIG. 1, except for the motion prediction process.

The input image signal is an image block having spatial and / or temporal motion prediction and motion compensation of the image, and having residual data based thereon. Is generated. The error data of the generated image block is transformed into coefficients by a 2-D transform method such as a discrete cosine transform (DCT) or a wavelet, and the transform coefficients are quantized in an arbitrary quantization step.

In case of predicting motion without considering the loss occurring in the quantization process, the decoder cannot restore the same video signal as the encoded video data. Accordingly, in consideration of this, motion compensation is performed by inverse quantization of the quantized transform coefficients and inverse transform, and a motion prediction operation is performed based on this.

The transform operation is performed to decorrelate the data generated by the motion prediction operation. A transform capable of inverse transform is applied. In addition, to reduce the hardware design of the encoder and decoder and the load of multiplication operations, integer transforms, 16-bit transforms, and 4x4 block size transforms instead of 8x8 are used. The 16x16 residual macro block is divided into 16 4x4 blocks, and each block is 4x4 integer converted.

The quantized transform coefficients are scanned in the same order as zigzag or modified zigzag, and then by entropy coding methods such as variable length coding (VLC) or arithmetic coding, It is coded into a compressed bit stream that is easy to transmit or store, along with motion vectors, header information, and additional information.

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. Accordingly, when providing a base layer, an interlayer prediction method may be used to increase coding efficiency.

When encoding a video block, the case in which the motion of the image is predicted and based on time (from a neighboring picture) is called inter mode. Otherwise, the case of using another pixel in the picture is referred to as intra mode. intra mode) or internal mode. In addition, a case in which an image block of an enhanced layer is predicted based on a block placed at a corresponding position in a frame of a simultaneous time of the base layer is called an intra base mode (intra base, intraBL, or intra interlayer mode).

Meanwhile, since the motion compensation prediction and transformation are performed in units of an image block, for example, a 16x16 macroblock or a 4x4 subblock, blocking artifacts occur at the boundary of the image block, thereby distorting the image. .

In order to reduce the blocking effect, deblocking filtering is applied after the motion prediction operation is performed. The deblocking filter smoothes the edges of the block to improve the quality of the decoded frame.

Deblocking filtering is applied to the vertical or horizontal edges of the image block, as shown in FIG. The deblocking filtering operation affects up to three pixels (samples) on both sides of the boundary. 2 shows four samples (p0, p1, p2, p3 and q0, q1, q2, q3) on both vertical or horizontal boundaries in adjacent image blocks p and q.

Whether or not deblocking filtering is applied to reduce block distortion is dependent on the intensity at the boundary of the image block and the gradient of pixels around the boundary, and the intensity of the deblocking filter is determined by the quantization parameter and the internal mode. , The type of the image block indicating the inter mode, the size of the block, the motion vector, the pixel value before deblocking filtering, and the like.

In general, a large value (3 or 4) of the boundary strength parameter (bs for short) is selected for blocks encoded in intra mode, and a small value (0 to 2 for non-intra mode blocks). Bs is selected.

In the intraBL mode, the deblocking filter may be applied to a picture (frame) of an up-sampled or interpolated base layer.

In the current Joint Scalable Video Model (JSVM) 1.0, in the case of intraBL mode, two deblocking filters are used after upsampling and after decoding, except in special cases. The case where the enhanced layer has the same resolution is included.

That is, when the spatial resolutions of the base layer and the enhanced layer are the same, the enhanced layer uses an image of the base layer that is interpolated without applying the deblocking filter as a reference image of intraBL prediction.

Also, for extended spatial scalability, for example, if the enhanced layer is a non-dyadic case where the spatial resolution is not an integer multiple of 2 of the base layer (de-blocking filter). There was not enough discussion about it.

The present invention has been made to solve the above problem, and an object of the present invention is to apply a deblocking filter even in the case of non-dyadic.

In addition, another object of the present invention is to reduce the number of conditions to be checked in order to apply the deblocking filter.

Another object of the present invention is to provide a method for appropriately adjusting the strength of the deblocking filter in the intraBL mode based on a base layer having the same spatial resolution.

In order to achieve the above object, a method of encoding / decoding a video signal according to an embodiment of the present invention includes: obtaining a second layer using the first layer as a reference video for prediction; And applying a second filter to the obtained second layer, wherein obtaining the second layer comprises: upsampling and / or interpolating any frame of the first layer; Applying a first filter to the upsampled and / or interpolated frames; And predicting a second layer based on the frame to which the first filter is applied, wherein the first filter and the second filter are filters for smoothing edges of the image block.

In the above embodiment, a filter having the same intensity is used as the first filter regardless of the resolution ratio of the first layer and the second layer, or when the second layer is twice the resolution of the first layer. Different strength filters can be used to distinguish the cases. It is also preferable that the first filter has a higher strength than the second filter.

Also, a method of encoding / decoding a video signal according to another embodiment of the present invention may include: obtaining a second layer using the first layer as a reference video for prediction; And applying a filter to smooth the edges of the image block to the obtained second layer, wherein the intensity of the filter is adjusted based on the resolution of the first layer and the second layer. It is done.

In the above embodiment, when the ratio of the resolution of the second layer and the first layer is not 2, it is preferable that a high intensity filter is applied.

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

In general, it is preferable to use an image block having a high intensity deblocking filter as a reference image for prediction in order to improve the efficiency of motion prediction. In addition, as described above, a high intensity deblocking filter is applied to the image block encoded in the intra mode.

As described above, when the enhanced layer uses the upsampled base layer as a reference image for prediction, the deblocking filter is applied after upsampling the base layer and after the decoding process (prediction operation).

In this case, namely, in the intraBL mode in which a base layer having a spatial resolution of 1/2 is used, the current JSVM applies a weak strength deblocking filter (bs = 1) to the image after the decoding process.

This is to prevent the image from becoming too dull by repeatedly applying a strong deblocking filter. Of course, a high intensity deblocking filter is applied to the base layer with a spatial resolution of 1/2 after upsampling.

To check whether the deblocking filter is applied to an upsampled or interpolated base layer, whether the enhanced layer and the base layer are dyadic, that is, the spatial resolution of the base layer is 1/2 of the spatial resolution of the enhanced layer. The process is necessary.

In a non-dyadic case including the case where the base layer and the enhanced layer have the same spatial resolution, the deblocking filter is not applied to the base layer image to be used as the reference image for intraBL prediction.

Therefore, in this case, even if the base layer is used as a reference image for the prediction of the enhanced layer, a problem arises in that a weak strength deblocking filter is applied only after the decoding process.

According to an embodiment of the present invention, a deblocking filter is always applied to a reconstructed image frame of a base layer and used for a reference image for inter-layer prediction of an enhanced layer. That is, the present invention uses a deblocking filter regardless of the ratio of the spatial resolution of the base layer and the enhanced layer.

In order to interpolate the reconstructed image frame of the base layer, a padding process and a deblocking process are performed. The deblocking filter is required even if the spatial resolution of the base layer and the enhanced layer is the same or dyadic. Therefore, it is necessary to adjust the strength of the deblocking filter in order to prevent the image from being over-smoothing by overlapping the deblocking filter.

In the case of the intra base layer, since the deblocking filter is used twice, after interpolation or upsampling, and after the decoding process, the high intensity deblocking filter may be too dull for the image. To prevent this, the strength of the deblocking filter must be adjusted.

In one embodiment of the present invention, the same deblocking filter is used in both the dyadic case and the non-dyadic case. In this case, there is no need to check whether dyadic / non-dyadic or not, but the strength of the deblocking filter is not adaptively adjusted.

In another embodiment of the present invention, a different deblocking filter is used to distinguish between the dyadic case and the non-dyadic case. In this case, it is necessary to check whether dyadic / non-dyadic, so the overall complexity increases, but it is possible to adaptively adjust the strength of the deblocking filter.

In particular, when the spatial resolutions of the base layer and the enhanced layer are the same, the reconstructed frame of the base layer is interpolated and a deblocking filter is applied to be used as a reference image of the enhanced layer in the intraBL mode.

In another embodiment of the present invention, the strength of the deblocking filter to be applied to the image after the decoding process is adjusted in consideration of the spatial resolutions of the base layer and the enhanced layer.

3 illustrates a process of selecting a boundary strength parameter bs when applying a deblocking filter to an image after a decoding process according to an embodiment of the present invention.

In the case of the deblocking filter after the decoding process, bs is generally selected according to the following rule for adjacent video blocks, for example, p and q of 4x4.

i) bs = 4 (strongest filtering) when p and / or q are intra mode and the boundary is the boundary of the macro block,

ii) when p and q are intra modes and the boundary is not the boundary of a macroblock, then bs = 3,

iii) both p and q are not in intra mode; when p or q include coded coefficients, bs = 2,

iv) neither p nor q is in intra mode; neither p nor q contain coded coefficients; When p and q refer to different frames, refer to different numbers of frames, or differ in the value of one luminance sample or more motion vector, bs = 1,

v) bs = 0 (not filtered) in all other cases

In the current JSVM, if the spatial resolution of the base layer is 1/2 of the enhanced layer, a high intensity (bs = 3 or 4) deblocking filter is applied after upsampling the base layer, and a low intensity after decoding. A deblocking filter of (bs = 1) is applied. When the spatial resolution of the base layer and the enhanced layer are the same or non-dyadic, the deblocking filter is not applied after the interpolation or upsampling with respect to the base layer, and the low blocking deblocking filter is applied even after the decoding process.

In one embodiment of the present invention, for p and q in intra mode or intra base mode and not at the boundaries of the macro block (in the same macro block, ie the modes of the two blocks are the same), p and q are intraBL and base Only when the spatial resolution of the layer is 1/2 of the enhanced layer, the strength of the deblocking filter after the decoding process is weakened (bs = 1), otherwise the strength of the deblocking filter after the decoding process is strengthened ( bs = 3).

Otherwise, this includes the case where p and q are intraBL and the spatial resolution of the base layer is not 1/2 of the enhanced layer and when p and q are intra mode (spatial resolution is meaningless in this case). .

In the former case, it is intra mode (or intra base mode), but since the deblocking filter is not applied after upsampling or interpolation, a high intensity (bs = 3) deblocking filter is applied after the decoding process. Since the latter case is an intra mode, a high intensity (bs = 3) deblocking filter is applied after the decoding process.

Also, for at least one of intra mode (intra mode or intra base mode) and p and q at the boundary of the macroblock, both p and q are intraBL and the spatial resolution of the base layer is 1/2 of the enhanced layer. Only the strength of the deblocking filter after the decoding process is weakened (bs = 1), otherwise the strength of the deblocking filter after the decoding process is strengthened (bs = 4).

Here, otherwise, if p and q are intraBL and the spatial resolution of the base layer is not 1/2 of the enhanced layer, if only one of p and q is intraBL, and if p or q is intra mode (In this case the spatial resolution is meaningless). Since all three cases are intra mode at the boundary of the macro block, a high intensity (bs = 4) deblocking filter is applied after the decoding process.

Thus, if the spatial resolution of the base layer and the enhanced layer is the same or non-dyadic, where no deblocking filter is applied after interpolation or upsampling on the base layer, then bs has a high intensity of 3 or 4 after the decoding process. A deblocking filter can be applied.

Basically, a deblocking filter is used in the process or apparatus for decoding an encoded picture. However, since the encoding apparatus must simulate the operation of the decoding apparatus for decoding the encoded bit stream in order to confirm the encoding performance, the encoding apparatus for encoding the video signal also performs the deblocking filtering operation. Accordingly, the present invention can be applied to a method / apparatus for encoding an image signal as well as a method / apparatus for decoding an encoded video signal.

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.

According to an embodiment of the present invention, the deblocking filter can be applied even in the case of extended spatial scalability, and the conventional problem that the deblocking filter is inappropriately weakly applied in the intraBL mode using a base layer having the same spatial resolution By resolving, we can compensate for the insufficient and incorrectly defined part of the current JSVM deblocking filter process.

Claims (6)

Decoding an image frame of a first layer; Applying a first deblocking filter to an image frame of the decoded first layer, and obtaining an image frame of the upsampled first layer; Decoding an image frame of a second layer using the image sample of the upsampled first layer; And And applying a second deblocking filter to the decoded image frame of the second layer. The method of claim 1, The first deblocking filter is a video signal decoding method, characterized in that the deblocking filter of the same intensity is used regardless of the resolution ratio of the first layer and the second layer. The method of claim 1, The second deblocking filter is a video signal decoding method characterized in that the deblocking filter of different intensities are applied to distinguish between the case where the resolution of the first layer and the second layer is the same and different. The method of claim 3, The second deblocking filter, if the resolution of the first layer and the second layer is different, the video signal decoding method characterized in that for applying a higher intensity deblocking filter than when the resolution is the same. Obtaining a second layer using the first layer as a reference image of the prediction; And Applying a deblocking filter to the obtained second layer, The intensity of the deblocking filter is adjusted based on the resolution of the first layer and the second layer. 6. The method of claim 5, And if the resolutions of the second layer and the first layer are different, a higher intensity deblocking filter is applied than if the resolution is the same.

Images