WO2014048374A1

WO2014048374A1 - Method and device for image processing

Info

Publication number: WO2014048374A1
Application number: PCT/CN2013/084531
Authority: WO
Inventors: 杨海涛; 杨名远; 张金雷; 李斌; 李厚强
Original assignee: 华为技术有限公司; 中国科学技术大学
Priority date: 2012-09-29
Filing date: 2013-09-27
Publication date: 2014-04-03
Also published as: CN103716622A; CN103716622B

Abstract

A method for image processing. The method comprises: according to a non-smooth area image block of an enhancement layer image and a non-smooth area image block of a base layer image, determining a first training filter, so that the first training filter satisfies: the degree of similarity between first forecasting information determined according to the first training filter and the non-smooth area image block of the base layer image and first original information determined according to the non-smooth area image block of the enhancement layer image satisfying a first preset condition, wherein the base layer image corresponds to the enhancement layer image; according to the first training filter, determining a candidate up-sampling filter, the candidate up-sampling filter comprising the first training filter; determining a target up-sampling filter from the candidate up-sampling filter; according to the target up-sampling filter and the base layer image block, determining forecasting information; and according to the forecasting information, performing coding processing on a target image block, so as to generate a target code stream.

Description

METHOD AND APPARATUS FOR IMAGE PROCESSING This application claims priority to Chinese Patent Application No. 201210374978.5, entitled "Method and Apparatus for Image Processing", filed on September 29, 2012, the The entire contents are incorporated herein by reference. Technical field

The present invention relates to the field of video processing and, more particularly, to a method and apparatus for image processing for image processing. Background technique

With the rapid development of the Internet and the increasing material and spiritual culture of people, the demand for video applications in the Internet, especially for high-definition video applications, is increasing, and the amount of high-definition video is very large. The problem that must be solved first in the transmission of limited bandwidth Internet is the high-definition video compression coding problem.

In a network environment (such as the Internet), since the network bandwidth is limited, the requirements of the terminal device and the user are different, so the code stream compressed once for a specific application is not satisfactory and effective. For some specific users or devices, it doesn't even make sense. An effective way to solve this problem is to use scalable video coding (SVC) techniques. In the SVC technique, an image is divided into a plurality of image layers based on quality parameters including spatial resolution, temporal resolution, or signal-to-noise ratio strength. The goal of SVC is to make the image layer with high quality make full use of the information of the image layer with low quality, and improve the efficiency of inter-layer prediction, so that the image with high quality can be more efficient.

When layering the image with high quality, because the resolution is different, it is necessary to upsample the image with low quality to achieve the resolution consistent with the image with high quality, and then use the information of the image with low quality to the enhancement layer. Make predictions. Therefore, the effect of upsampling has a direct impact on the coding performance. Due to the different characteristics of the texture of the image, the effect of upsampling can be effectively improved by using multiple filters.

At present, a technique of performing upsampling using a plurality of filters is known, which calculates the horizontal direction and the vertical direction of four pixels of positions 1, 2, 3, 4 for each 4 x 4 image block. Step difference, adding the second-order difference in the horizontal direction of four pixels as a horizontal (horizontal), will be four The second-order differential addition of the pixels in the vertical direction is as vertical. Thereby, the texture direction can be determined. Example: 3⁄4 mouth,: 3⁄4 mouth fruit is twice larger than horizontal, then the direction is 1; ^ mouth fruit horizontal is greater than twice, and 'J determines direction is 2; ^ mouth fruit is equal to twice horizontal Bay ¹ J determines that direction is 0. Thereafter, the average values of vertical and horizontal are calculated to determine the type of the pixel block, and a filter is separately trained for each type, so that the trained filter is very consistent with the corresponding category, but the method needs to be encoded. The end encodes the coefficients of the filter and the index of the filter used by each type of block (index) and transmits it to the decoder so that it can be decoded correctly at the decoder. If this method is used to upsample the base layer reconstructed image, it will result in too many bits representing the filter coefficients and index to be transmitted, thereby affecting the encoding performance. Summary of the invention

Embodiments of the present invention provide a method and apparatus for image processing, which can improve the effect and performance of image processing.

In a first aspect, a method for image processing is provided, the method comprising: determining a first training filter according to a non-smooth area image block of an enhancement layer image and a non-smooth area image block of a base layer image, so that the The first training filter satisfies: first prediction information determined according to the first training filter and the non-smooth area image block of the base layer image and first original information determined according to the non-smooth area image block of the enhancement layer image The similarity between the two meets a first preset condition, wherein the base layer image corresponds to the enhancement layer image; and according to the first training filter, an alternative upsampling filter is determined, wherein the candidate upsampling filter The first training filter is included; from the candidate upsampling filter, the target upsampling filter is determined; the prediction information is determined according to the target upsampling filter and the base layer image block; and the target is determined according to the prediction information The image block is subjected to an encoding process to generate a target code stream, wherein the target image block is located in the enhancement layer image, the base Layer of the image block positioned base layer image, and the spatial position of the block in the base image in a base layer image to the target image block in the spatial position of the image enhancement layer, respectively.

In a possible implementation, the method further includes: determining, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter Satisfying: satisfying the second degree of similarity between the second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image Pre-set conditions, wherein the base layer image and the increase The strong layer image corresponds to; and determining the candidate upsampling filter according to the first training filter, further comprising: determining an alternative upsampling filter according to the first training filter and the second training filter, The candidate upsampling filter includes the first training filter and the second training filter.

With reference to the first aspect and the first possible implementation manner, in the second possible implementation manner, the determining the target upsampling filter from the candidate upsampling filter comprises: according to characteristics of the base layer image block Determining the smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or residual information of the base layer image block; determining the target according to the smoothness of the target image block Upsampling filter.

With reference to the first aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, determining, according to the first training filter and the second training filter, The sampling filter further includes: determining an alternative upsampling filter according to the first training filter, the second training filter, and the conventional filter, wherein the candidate upsampling filter includes the first training filter And the second training filter and the conventional filter; and the encoding processing the target image block to generate the target code stream according to the prediction information, comprising: encoding the target image block according to the prediction information And generating a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, where the first indication information is used to obtain the target upsampling when decoding the encoded target image block. The basis of the filter.

With reference to the first aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in a fourth possible implementation manner, the determining, according to the first training filter, Selecting the sampling filter further includes: determining an alternative upsampling filter according to the first training filter and the conventional filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter And determining the target upsampling filter, comprising: determining a texture degree of the target image block; determining the target upsampling filter according to the texture degree of the target image block.

In a second aspect, a method for image processing is provided, the method comprising: determining a first training filter according to a non-smooth area image block of an enhancement layer image and a non-smooth area image block of a base layer image, so that the The first training filter satisfies: first prediction information determined according to the first training filter and the non-smooth area image block of the base layer image and first original information determined according to the non-smooth area image block of the enhancement layer image The similarity between the two meets a first preset condition, wherein the base layer image corresponds to the enhancement layer image; and according to the first training filter, an alternative upsampling filter is determined, wherein the candidate upsampling filter The first training filter is included; In the sampling filter, a target upsampling filter is determined, the candidate upsampling filter including the first training filter; determining prediction information according to the target upsampling filter and the base layer image block; and based on the prediction information and the slave Residual information obtained in the target code stream, the target code stream is decoded to obtain the target image block, wherein the target image block is located in the enhancement layer image, and the base layer image block is located in the base layer image. And the spatial position of the basic image block in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image.

In a possible implementation, the method further includes: determining, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter Satisfying: satisfying the second degree of similarity between the second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image a preset condition, wherein the base layer image corresponds to the enhancement layer image; and determining the candidate upsampling filter according to the first training filter, further comprising: according to the first training filter and the first And a second training filter, the candidate upsampling filter is determined, wherein the candidate upsampling filter comprises the first training filter and the second training filter.

With reference to the second aspect and the first possible implementation manner, in a second possible implementation manner, in the alternative upsampling filter, the candidate upsampling filter is used to determine a target upsampling filter, The method includes: determining, according to the feature information of the base layer image block, a smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or residual information of the base layer image block; The smoothness of the target image block determines the target upsampling filter.

With reference to the second aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, determining, according to the first training filter and the second training filter, The sampling filter further includes: determining an alternative upsampling filter according to the first training filter, the second training filter, and the conventional filter, wherein the candidate upsampling filter includes the first training filter And the second training filter and the conventional filter; and the candidate upsampling filter, determining the target upsampling filter, comprising: obtaining, from the target code stream, a target upsampling filter The first indication information is determined according to the first indication information, and the target upsampling filter is determined.

With reference to the second aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in the fourth possible implementation manner, the determining, according to the first training filter, Selecting the sampling filter further includes: according to the first training filter and the conventional filter, Determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; and the determining the target upsampling filter, comprising: determining a texture degree of the target image block; The target upsampling filter is determined according to the texture of the target image block.

In a third aspect, an apparatus for image processing is provided, the apparatus comprising: an obtaining unit, configured to determine a first training filter according to a non-smooth area image block of an enhancement layer image and a non-smooth area image block of a base layer image So that the first training filter satisfies: the first prediction information determined according to the first training filter and the non-smooth area image block of the base layer image and the non-smooth area image block according to the enhancement layer image The similarity between the first original information satisfies a first preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit determines an alternative upsampling filter according to the first training filter, The candidate upsampling filter includes the first training filter; and the target upsampling filter is determined from the candidate upsampling filter; the encoding unit is configured to obtain the upsampling filter from the determining unit And determining prediction information according to the target upsampling filter and the base layer image block; for using the prediction information, The image block is subjected to an encoding process to generate a target code stream, wherein the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and a space of the basic image block in the base layer image is The position corresponds to a spatial position of the target image block in the enhancement layer image.

In a possible implementation, the acquiring unit is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter The device satisfies: the similarity between the second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit is further configured to determine an candidate upsampling filter according to the first training filter and the second training filter, The candidate upsampling filter includes the first training filter and the second training filter.

With reference to the third aspect and the first possible implementation manner, in a second possible implementation manner, the determining unit is specifically configured to determine, according to feature information of the base layer image block, a smoothness of the target image, where The feature information includes coded block flag information of the base layer image block or residual information of the base layer image block; and is used to determine the target upsampling filter according to the smoothness of the target image block.

With reference to the third aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, the determining unit is further configured to: according to the first training filter, a second training filter and a conventional filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter, the second training filter, and the conventional filter; and the encoding The unit is specifically configured to perform coding processing on the target image block according to the prediction information, to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, where the first indication information is used by It is used as a basis for acquiring the target upsampling filter when decoding the above-mentioned encoded target image block.

With reference to the third aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in a fourth possible implementation manner, the determining unit is further configured to perform, according to the first training a filter and a conventional filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; specifically for determining a texture degree of the target image block; The target upsampling filter is determined according to the texture degree of the target image block.

According to a fourth aspect, an apparatus for image processing is provided, the apparatus comprising: an obtaining unit, configured to determine, according to a non-smooth area image block of an enhancement layer image and a non-smooth area image block of a base layer image, a first training filter, so that the first training filter satisfies: first prediction information determined according to the first training filter and the non-smooth area image block of the base layer image and a non-smooth region according to the enhancement layer image The similarity between the first original information determined by the image block satisfies a first preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit is configured to determine, according to the first training filter, Selecting a sampling filter, wherein the candidate upsampling filter comprises the first training filter; and from the candidate upsampling filter, determining a target upsampling filter, the candidate upsampling filter comprising the a first training filter; a decoding unit, configured to acquire the target upsampling filter from the determining unit, and perform upsampling filtering according to the target Determining prediction information according to the base layer image block, and performing decoding processing on the target code stream according to the prediction information and the residual information obtained from the target code stream, where the target image block is obtained, where the target image block is obtained Located in the enhancement layer image, the base layer image block is located in the base layer image, and a spatial position of the base image block in the base layer image corresponds to a spatial position of the target image block in the enhancement layer image.

In a possible implementation, the acquiring unit is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter The device satisfies: the second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the smooth region image block according to the enhancement layer image The similarity between the second original information satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit is further configured to use the first training filter and the second training And a filter determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the second training filter.

With reference to the fourth aspect and the first possible implementation manner, in a second possible implementation manner, the determining unit is specifically configured to determine, according to feature information of the base layer image block, a smoothness of the target image, where The feature information includes coded block flag information of the base layer image block or residual information for the base layer image block; and the target upsampling filter is determined according to the smoothness of the target image block.

With reference to the fourth aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, the determining unit is further configured to perform, according to the first training filter, the second training filter And a conventional filter, the candidate upsampling filter is determined, wherein the candidate upsampling filter comprises the first training filter, the second training filter, and the conventional filter; And obtaining first indication information for indicating the target upsampling filter, and determining, according to the first indication information, the target upsampling filter.

With reference to the fourth aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in a fourth possible implementation manner, the determining unit is further configured to perform, according to the first training a filter and a conventional filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; for determining a texture degree of the target image block; The target upsampling filter is determined according to the texture of the target image block.

In a fifth aspect, an encoder for image processing is provided, the encoder comprising: a bus; a processor coupled to the bus; a memory coupled to the bus; wherein the processor calls the memory through the bus a program stored in the method for determining a first training filter according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, so that the first training filter satisfies: according to the first A similarity between the first prediction information determined by the training filter and the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, Wherein the base layer image corresponds to the enhancement layer image; determining an alternative upsampling filter according to the first training filter, wherein the candidate upsampling filter comprises the first training filter; In the alternative upsampling filter, determining a target upsampling filter; determining a prediction based on the target upsampling filter and the base layer image block Information; encoding, according to the prediction information, a target image block to generate a target code stream, where The target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and a spatial position of the basic image block in the base layer image and a space of the target image block in the enhancement layer image The location corresponds.

In a possible implementation, the processing unit is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter The device satisfies: the similarity between the second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a preset condition, wherein the base layer image corresponds to the enhancement layer image; and configured to determine an alternative upsampling filter according to the first training filter and the second training filter, where the candidate The upsampling filter includes the first training filter and the second training filter.

With reference to the fifth aspect and the first possible implementation manner, in a second possible implementation manner, the processing unit is specifically configured to determine, according to feature information of the base layer image block, a smoothness of the target image, where The feature information includes coded block flag information of the base layer image block or residual information of the base layer image block; and is used to determine the target upsampling filter according to the smoothness of the target image block.

With reference to the fifth aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, the processor is specifically configured to perform, according to the first training filter, the second training filter And an alternative filter, wherein the candidate upsampling filter comprises the first training filter, the second training filter and the conventional filter; and based on the prediction information, Encoding the target image block to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, the first indication information is used to decode the encoded target image Block as the basis for obtaining the target upsampling filter

With reference to the fifth aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in a fourth possible implementation manner, the processor is specifically configured to use the first training a filter and a conventional filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; specifically for determining a texture degree of the target image block; The target upsampling filter is determined according to the texture degree of the target image block. .

In a sixth aspect, a decoder for image processing is provided, the decoder comprising: a bus; a processor coupled to the bus; a memory coupled to the bus; wherein the processor calls the memory through the bus a program stored in the image for use in a non-smooth area image of the enhancement layer image Determining, by the non-smooth area image block of the block and the base layer image, the first training filter, so that the first training filter satisfies: determining according to the first training filter and the non-smooth area image block of the base layer image A similarity between a prediction information and a first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein the base layer image corresponds to the enhancement layer image; a first training filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter; and from the candidate upsampling filter, determining a target upsampling filter, the candidate The upsampling filter includes the first training filter; determining prediction information according to the target upsampling filter and the base layer image block; and according to the prediction information and the residual information obtained from the target code stream, the target code stream Decoding processing to obtain the target image block, wherein the target image block is located in the enhancement layer image, and the base layer image block is located at the base The spatial position of the tomographic image, and the spatial position of the image block is substantially in the base layer image in the target image block in the enhancement layer corresponding to the image.

With reference to the sixth aspect and the first possible implementation manner, in a second possible implementation manner, the processing unit is specifically configured to determine, according to feature information of the base layer image block, a smoothness of the target image, where The feature information includes coded block flag information of the base layer image block or residual information for the base layer image block; and the target upsampling filter is determined according to the smoothness of the target image block.

With reference to the sixth aspect, the first possible implementation manner, and the second possible implementation manner, in a third possible implementation manner, the processor is further configured to perform, according to the first training filter, the second training filter And a conventional filter, the candidate upsampling filter is determined, wherein the candidate upsampling filter comprises the first training filter, the second training filter, and the conventional filter; And obtaining first indication information for indicating the target upsampling filter, and determining, according to the first indication information, the target upsampling filter. With reference to the sixth aspect, the first possible implementation manner, the second possible implementation manner, and the third possible implementation manner, in a fourth possible implementation manner, the processor is further configured to perform, according to the first training a filter and a conventional filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; for determining a texture degree of the target image block; The target upsampling filter is determined according to the texture of the target image block.

A method, an apparatus encoder, and a decoder for image processing according to an embodiment of the present invention can reduce an upsampling filter while improving an upsampling effect by determining an upsampling filter according to a smooth region and a non-smooth region. The number does not need to transmit filter coefficients and indexes, so that the encoding performance can be improved, and the effect and performance of image processing can be improved. DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings to be used in the embodiments of the present invention will be briefly described below. Obviously, the drawings described below are only some embodiments of the present invention. Other drawings may also be obtained from those of ordinary skill in the art in view of the drawings.

1 is a schematic flow chart of a method for image processing according to an embodiment of the present invention. 2 is a schematic diagram of a method for determining a training filter, in accordance with an embodiment of the present invention. 3 is another schematic flow diagram of a method for image processing in accordance with an embodiment of the present invention.

4 is a schematic block diagram of an apparatus for image processing in accordance with an embodiment of the present invention. FIG. 5 is a schematic block diagram of an apparatus for image processing according to another embodiment of the present invention. 6 is a schematic block diagram of an encoder for image processing in accordance with an embodiment of the present invention. FIG. 7 is a schematic block diagram of a decoder for image processing according to another embodiment of the present invention. detailed description

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without making creative labor are within the scope of the present invention.

FIG. 1 shows a schematic flow diagram of a method 100 for image processing in accordance with an embodiment of the present invention as described from the perspective of an encoding end. As shown in FIG. 1, the method 100 includes: S110. Determine, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter, so that the first training filter satisfies: according to the first training filter and the basic The similarity between the first prediction information determined by the non-smooth area image block of the layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein the basic layer The image corresponds to the enhancement layer image;

S120. Determine, according to the first training filter, an alternative upsampling filter, where the candidate upsampling filter includes the first training filter;

S130. Determine, from the candidate upsampling filter, a target upsampling filter;

S140. Determine prediction information according to the target upsampling filter and the base layer image block.

S150. Perform encoding processing on the target image block according to the prediction information, to generate a target code stream, where the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and the basic image is The spatial position of the block in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image

In the layered encoding of the image, for example, in spatial scalable coding, the image can be processed to obtain a low resolution image, and the original image is referred to as a high resolution image as a contrast, and the encoder separately determines the low resolution. The rate image and the high resolution image are encoded. For convenience of description, a high quality image to be encoded is referred to herein as an enhancement layer image, and a corresponding low quality image to be encoded (e.g., the low resolution image) is referred to as a base layer image.

In the embodiment of the present invention, the target image is an image processed by a layered coding technique, and the basic layer refers to a quality in layered coding (including frame rate, spatial resolution, temporal resolution, signal to noise ratio intensity or quality level). The lower layer, the enhancement layer refers to the layer with higher quality (including frame rate, spatial resolution, temporal resolution, signal-to-noise ratio intensity or quality level) in the layered coding. It should be noted that, in the embodiment of the present invention, in the embodiment of the present invention, for a given enhancement layer, the corresponding base layer may be any layer lower in quality than the enhancement layer, for example, if currently There are five layers, and the coding quality is sequentially improved (that is, the first layer has the lowest quality and the fifth layer has the highest quality). If the enhancement layer is the fourth layer, the base layer may be the first layer or the second layer. It is the third layer or the fourth layer. Similarly, for a given base layer, the corresponding enhancement layer can be any layer of lower quality than the base layer.

The enhancement layer image is the image in the currently processed enhancement layer, and the base layer image is the image in the base layer at the same time as the enhancement layer image.

In summary, in the embodiment of the present invention, the quality of the base layer image is lower than the enhancement layer image. the quality of.

The target image block is the image block being processed in the enhancement layer image.

The base layer image block is an image block in the base layer image that has a corresponding relationship with the target image block in spatial position.

In the embodiment of the present invention, the correspondence between the base layer image block and the target image block may be calculated according to the resolution proportional relationship between the base layer image and the enhancement layer image. For example, in a system including the X direction and the y direction, if the resolution of the enhancement layer image in the X direction and the y direction is twice that of the base layer image, respectively, the pixel coordinates of the upper left corner in the enhancement layer are (2x, 2y). And an image block of size (2m) X ( 2n ), the corresponding block in the base layer image may be an image block whose pixel coordinates are (X , y ) and whose size is mxn in the upper left corner.

In the embodiment of the present invention, an image (including an enhancement layer image and a base layer image) is divided into two types of regions, that is, a smooth region and a non-smooth region. In the smoothing region, a more accurate prediction value is usually obtained during encoding, so that the coded block flag (CBF, Coded Block Flag) determined at the final encoding is 0, that is, the residual is 0. In the non-smooth area, the signal characteristics are not highly correlated during encoding, so it cannot be accurately predicted. Therefore, the more accurate prediction values cannot be obtained, resulting in a larger or uneven residual after prediction. Therefore, the final coded CBF is not zero.

Therefore, in the embodiment of the present invention, it may be determined that an image block with a CBF of 0 in the base layer image (a smooth region image block of the base layer image) is located in the smooth region, and thus, may be considered as an enhancement layer image corresponding to the image block. The image block (the smooth area image block of the enhancement layer image) is located in the smooth area. Similarly, it can be determined that the image block in which the CBF is not 0 in the base layer image (the non-smooth area image block of the base layer image) is located in the non-smooth region, and thus, the image block in the enhancement layer image corresponding to the image block can be considered (The non-smooth area image block of the enhancement layer image) is located in the non-smooth area. As described above, the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, and the non-smooth region image block of the enhancement layer image and the non-smooth region image block of the base layer image can be determined.

Thereafter, at S110, the first training filter may be trained according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image. As an example and not by limitation, the first method may be trained by the following method. Training filter.

As shown in FIG. 2, yl(n) represents a smooth region image block of an enhancement layer image (original image) (a plurality of image fast collections, hereinafter referred to as original smooth region image blocks), and x(n) represents a base layer image. Smooth region image block, H(n) represents the filter to be determined (first training filter), e(n) represents A smooth region image block (referred to as a predicted image block) of the base layer image after filtering the wave and the original smooth region image block. For example, to minimize the mean square error of the pixel values of the original smooth region image block and the predicted image block, as an optimization condition (an example of the first preset condition), H(n) is solved.

The optimal solution of H(n) can be expressed as the following formula (1),

H = R ¹ xr ( 1 )

Where R _xx represents the autocorrelation function of the original smooth region image block, and r _yx represents the cross-correlation function of the smooth region image block of the base layer image and the original smooth region image block.

It should be understood that the above-described methods of acquiring the training filter are merely illustrative, and the present invention is not limited thereto. For example, the method of acquiring the training filter can be variously changed by changing the above-described optimization conditions and the like.

Optionally, the method further includes:

Determining, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter to satisfy the second training filter: according to the second training filter and the base layer image The similarity between the second prediction information determined by the smooth region image block and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a second preset condition, wherein the base layer image and the enhancement Layer image corresponds; and

Determining the candidate upsampling filter according to the first training filter, further comprising: determining, according to the first training filter and the second training filter, an alternative upsampling filter, wherein the candidate upsampling The filter includes the first training filter and the second training filter.

In the embodiment of the present invention, the training samples are different except that the first training filter is a non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image is obtained, and the second training filter is the enhancement layer. The smooth region image block of the image and the smooth region image block of the base layer image are obtained by training, and the other processes are similar. Here, in order to avoid redundancy, the description thereof is omitted.

Through the foregoing method, the first training filter and the second training filter can be respectively obtained. Therefore, in the embodiment of the present invention, the coefficients of the training filter are fixed, and the bit transmission filter coefficients are not consumed in the encoding process. When the filter is selected at the decoding end, for example, the type of the current block (ie, a smooth area image block or a non-smooth area image block) may be determined according to the CBF value of the corresponding block of the base layer, thereby eliminating the need to transmit each image block. The index of the filter improves the performance of image processing.

Thus, the first filter and the second training filter may be included in the alternative filter, ie, case 1. Situation 1

Optionally, determining the target upsampling filter from the candidate upsampling filter, comprising: determining a smoothness of the target image according to the feature information of the base layer image block, where the feature information includes the base layer Encoding block flag information of an image block or residual information for the base layer image block;

The target upsampling filter is determined according to the smoothness of the target image block.

Specifically, in the case where the candidate upsampling filter includes the first training filter and the second training filter acquired as described above, in S130, according to the type of the target image block (for the smooth region image block) Or non-smooth area image block), select the target upsampling filter.

In the embodiment of the present invention, when the CBF of the base layer image block corresponding to the target image block is 0, it may be determined that the target image block is located in the smooth region. Similarly, when the CBF of the base layer image block corresponding to the target image block is not 0, it can be determined that the target image block is located in the non-smooth region.

It should be understood that the above-listed methods for determining the smoothness of the target image block and the parameters used are merely illustrative, and the present invention is not limited thereto. Others can determine the smoothness of the target image block, that is, determine that the target image block is in Methods and parameters for smoothing or non-smoothing regions in an enhancement layer image fall within the scope of the present invention.

In this case, at S140, the reconstructed image of the base layer image block may be upsampled according to the target upsampling filter determined as described above, thereby determining the predicted value with respect to the target image block.

Thus, at S150, the distortion rate cost of each prediction mode can be calculated, and the optimal prediction mode used when encoding the target image block is determined. It should be noted that, in the decoding process of the encoded target image block, the target upsampling filtering crying port can be determined in the same manner as described above.

Optionally, the candidate upsampling filter further includes a conventional filter, and

The encoding process of the target image block to generate the target code stream according to the prediction information includes:

And encoding, according to the prediction information, the target image block to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, where the first indication information is used to decode the foregoing The encoded target image block is used as the basis for acquiring the target upsampling filter. In the embodiment of the present invention, as a conventional filter, for example, an existing cosine interpolation filter may be included. In the following description, for convenience of explanation, the conventional filter is taken as an example of a cosine interpolation filter.

Specifically, in the embodiment of the present invention, the foregoing method (training process) for acquiring a training filter is performed in a pixel domain, and may be different for a downsampling ratio (a resolution ratio between an enhancement layer image and a base layer image) Different training is performed, for example, there is a case where the filter obtained for the entire pixel position and the half pixel position is different, and therefore, the prediction of the integer pixel position and the half pixel position may be unbalanced, thereby affecting the smoothness of the prediction residual. Therefore, in the embodiment of the present invention, a conventional filter, such as a cosine interpolation filter, may be added to the candidate upsampling filter, and the method for acquiring the cosine interpolation filter may be the same as the prior art. Here, in order to avoid redundancy, the description thereof is omitted.

Thus, the alternative filter may include the first training filter and the second training filter and a cosine interpolation filter (an example of a conventional filter), that is, Case 2.

Situation 2

In S 130, for example, rate distortion optimization (RDO, Rate-Distortion) can be used.

Optimization ), choose to use the cosine interpolation filter or the first training filter or the second training filter as the target upsampling filter.

In this case, thereafter, at S140, the reconstructed image of the base layer image block may be upsampled based on the target upsampling filter to determine a predicted value with respect to the target image block.

Thus, at S150, the distortion rate cost of each prediction mode can be calculated, and the optimal prediction mode used when encoding the target image block is determined. In the case that the optimal prediction mode is the inter-layer texture prediction mode, it may be used to indicate the target upsampling filter (specifically, indicating that the target upsampling filter is a cosine interpolation filter, a first training filter) Or the first indication information of the filter of the first training filter is encoded, and the first instruction information after the encoding process is added to the code stream to obtain the target when decoding the encoded target image block. Upsampling filter basis. It should be noted that, when the candidate filter includes the cosine interpolation filter or the first training filter or the second training filter as the target upsampling filter, the candidate filters are three, and, as described above, The encoding end or the decoding end may determine to use the first training filter or the second training filter according to the smoothness of the target image block, and therefore, the first indication information may be carried by using only one identifier in the code stream, Indicates whether the target upsampling filter is a conventional filter (cosine interpolation filter) or a training filter (first training filter or second training filter). Alternative filtering In the device, when the number of the conventional filters is at least two, the purpose of indicating the target image block can be achieved by increasing the number of bits of the first indication information.

When upsampling by the cosine interpolation filter, the corresponding pixel points of the base layer image are always directly copied. However, in the scalable video coding method, for example, the pixels of the base layer image are not directly copied by the corresponding enhancement layer image. Pixel, at this time, if the whole pixel of the target image is located at the abrupt point, directly copying the corresponding pixel of the base layer during upsampling will cause a large error. Therefore, using this method for upsampling may result in a large error in the entire pixel position of the enhancement layer image. Therefore, in the embodiment of the present invention, the pixel of the above-mentioned mutation point is trained by using the method of the embodiment of the present invention. The training filter (the first training filter or the second training filter) can improve the effect of upsampling and improve the performance of encoding and image processing.

The determining target upsampling filter comprises:

Determining the texture of the target image block;

The target upsampling filter is determined according to the texture of the target image block.

As described above, in the practice of the present invention, in the case where the alternative upsampling filter comprises a cosine interpolation filter, a first training filter and a second training filter. As described above, after selecting the target upsampling filter according to the RDO, it is necessary to transmit a flag bit for carrying the encoded first indication information in the code stream to determine the target upsampling when decoding the encoded target image block. Which filter is the filter. Therefore, two candidate filters can be used in this scheme, but they are not selected by RDO, but are selected according to different characteristics of different filters, so as to avoid multi-transmission of one marker bit to represent the target up-sampling filter. That is, the alternative filter includes a first training filter and a conventional filter, that is, case 3

Situation 3

In S130, for example, the target upsampling filter may be selected from the first training filter and the cosine interpolation filter (an example of a conventional filter) according to the feature information of the base layer image block, thereby avoiding the use of the transmission flag bit representation. The filter may be determined, for example, by using the texture strength of the base layer image block. In the embodiment of the present invention, for example, edge detection, texture strength analysis, or intra-frame coding prediction mode may be used. The texture of the base layer image block is determined, and, for example, when the texture is strong, the training-derived filter is selected as the target up-sampling filter, and the rest uses the cosine interpolation filter. The above enumerated methods for determining the texture of the base layer image block by edge detection, texture intensity analysis or intra-frame coding prediction mode The method can be the same as the prior art, and the description thereof will be omitted herein in order to avoid redundancy. In this case, thereafter, at S140, the reconstructed image of the base layer image block may be upsampled according to the target upsampling filter to determine a predicted value with respect to the target image block.

Thus, at S150, the distortion rate cost of each prediction mode can be calculated, and the optimal prediction mode used when encoding the target image block is determined. The target image block is encoded.

When upsampling by the cosine interpolation filter, the corresponding pixel points of the base layer image are always directly copied. However, in the scalable video coding method, for example, the pixels of the base layer image are not directly copied by the corresponding enhancement layer image. Pixel, at this time, if the whole pixel of the target image is located at the abrupt point, directly copying the corresponding pixel of the base layer during upsampling will cause a large error. The first training filter can better recover the pixels at the position of the mutation point, so it can compensate for the deficiency of the cosine interpolation filter.

It should be noted that, when decoding the encoded target image block, the target upsampling filter can be determined in the same manner as described above.

According to the method for image processing according to the embodiment of the present invention, by determining the upsampling filter according to the smoothed region and the non-smoothed region, the number of upsampling filters can be reduced and the transmission filter can be eliminated while improving the effect of upsampling Coefficients and indexes, which can improve the coding performance and improve the effect and performance of image processing.

Fig. 3 shows a schematic flow chart of a method 200 for image processing according to an embodiment of the present invention, which is described from the perspective of an encoding end. As shown in FIG. 3, the method 200 includes:

S210. Determine, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter, so that the first training filter satisfies: according to the first training filter and the basic The similarity between the first prediction information determined by the non-smooth area image block of the layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein the basic layer The image corresponds to the enhancement layer image;

S220, determining, according to the first training filter, an alternative upsampling filter, where the candidate upsampling filter includes the first training filter;

S230. Determine, from an alternative upsampling filter, a target upsampling filter, where the candidate upsampling filter includes the first training filter;

S240. Determine prediction information according to the target upsampling filter and the base layer image block. S250. Decode the target code stream according to the prediction information and the residual information obtained from the target code stream, to obtain the target image block, where the target image block is located in the enhancement layer map. In the image, the base layer image block is located in the base layer image, and a spatial position of the base image block in the base layer image corresponds to a spatial position of the target image block in the enhancement layer image.

In summary, in the embodiment of the present invention, the quality of the base layer image is lower than the quality of the enhancement layer image.

In the embodiment of the present invention, the image (including the enhancement layer image and the base layer image) is divided into a single sheet Two types of areas, namely smooth areas and non-smooth areas. In the smooth region, a more accurate prediction value can usually be obtained during encoding, so that the coded block flag (CBF, Coded Block Flag) determined at the final encoding is 0, that is, the residual is 0. In the non-smooth region, the signal characteristics are not highly correlated during encoding, so it is impossible to accurately predict, so that a more accurate prediction value cannot be obtained, resulting in a large or uneven residual after prediction, and thus the final encoded CBF is not zero.

Thereafter, in S210, the first training filter may be trained according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image. As an example and not by limitation, the first method may be trained by the following method. Training filter.

As shown in FIG. 2, yl(n) represents a smooth region image block of an enhancement layer image (original image) (a plurality of image fast collections, hereinafter referred to as original smooth region image blocks), and x(n) represents a base layer image. Smooth area image block, H(n) represents the filter to be determined (first training filter), e(n) represents the smoothed area image block of the base layer image after filtering (denoted as prediction image block) and original Smooth area image block. For example, the mean square error of the pixel values of the original smooth region image block and the predicted image block is minimized, and H(n) is solved as an optimization condition (an example of the first preset condition).

The optimal solution of H(n) can be expressed as the above formula (1),

Optionally, the method further includes:

Determining, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter to satisfy the second training filter: according to the second training filter and the base layer image The second prediction information determined by the smooth region image block and according to the enhancement The similarity between the second original information determined by the smooth region image block of the layer image satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image;

Thus, the alternative filter may include the first training filter and the second training filter, ie, Case 4.

Situation 4

Specifically, in the case where the candidate upsampling filter includes the first training filter and the second training filter acquired as described above, in S230, according to the type of the target image block (for the smooth region image block) Or non-smooth area image block), select the target upsampling filter.

It should be understood that the above enumerated methods for determining the smoothness of the target image block and the parameters used are only Illustratively, the present invention is not limited thereto, and other methods and parameters capable of determining the smoothness of a target image block, that is, determining a smooth region or a non-smooth region of the target image block in the enhancement layer image, fall within the present invention. The scope of protection.

In this case, at S240, the reconstructed image of the base layer image block may be upsampled according to the target upsampling filter determined as described above, thereby determining a predicted value with respect to the target image block.

Therefore, in S250, the target image block may be decoded according to the predicted value and the residual information obtained from the target code stream. In the embodiment of the present invention, the process may be the same as the prior art. The description is omitted.

In the alternative upsampling filter, the target upsampling filter is determined, including:

And obtaining, from the target code stream, first indication information for indicating the target upsampling filter, and determining the target upsampling filter according to the first indication information.

In the embodiment of the present invention, as a conventional filter, for example, an existing cosine interpolation filter may be included. In the following description, for convenience of explanation, the conventional filter is taken as an example of a cosine interpolation filter.

Thus, the alternative filter may include the first training filter and the second training filter and the cosine interpolation filter (an example of a conventional filter), that is, case 5.

Situation 5

In S230, for example, the first indication information may be obtained from the target code stream. In the embodiment of the present invention, the candidate filter includes the cosine interpolation filter or the first training filter or the second training filter, as When the target upsampling filter, the candidate filter is three, and, as described above, the encoding end or the decoding end may determine whether to use the first training filter or the first according to the smoothness of the target image block. a second training filter, therefore, the first indication information can be carried in the code stream using only one bit identifier to indicate whether the target upsampling filter is a conventional filter (cosine interpolation filter) or a training filter (first a training filter or a second training filter). For example, when the indication information is 1, selecting a training filter (including a first training filter or a second training filter), and according to the smoothness of the target image block, from the first training filter or the second training filter The target upsampling filter is determined in the device. For example, when the indication information is 0, the cosine interpolation filter is selected as the target upsampling filter.

It should be understood that the above-described method of indicating the target upsampling filter by the first indication information is merely illustrative, and the present invention is not limited thereto. The purpose of indicating the target image block is achieved by increasing the number of bits of the first indication information.

The determining target upsampling filter comprises:

Determining the texture of the target image block;

As described above, in the practice of the present invention, in the case where the alternative upsampling filter includes a cosine interpolation filter, a first training filter, and a second training filter. As described above, after selecting the target upsampling filter according to the RDO, the first indication signal for carrying the encoding needs to be transmitted in the code stream. The flag bit of the bit is used to determine which filter the target upsampling filter is when decoding the above-described encoded target image block. Therefore, two candidate filters can be used in this scheme, but it does not pass.

RDO selects, but chooses according to the different characteristics of different filters, so as to avoid multi-transmission of one marker bit to represent the target upsampling filter. That is, the alternative filter includes a first training filter and a conventional filter, that is, case 6

Situation 6

In S230, for example, a target upsampling filter may be selected from the first training filter and the cosine interpolation filter (an example of a conventional filter) according to the feature information of the base layer image block, thereby avoiding transmission of the flag bit representation used. The filter may be determined, for example, by using the texture strength of the base layer image block. In the embodiment of the present invention, for example, edge detection, texture strength analysis, or intra-frame coding prediction mode may be used. The texture of the base layer image block is determined, and, for example, when the texture is strong, the training-derived filter is selected as the target up-sampling filter, and the rest uses the cosine interpolation filter. The method for determining the texture of the base layer image block by the method of edge detection, texture intensity analysis or intraframe coding prediction mode, which is enumerated above, may be the same as the prior art, and the description thereof will be omitted herein to avoid redundancy.

According to the method for image processing according to the embodiment of the present invention, by determining the upsampling filter according to the smoothed region and the non-smoothed region, the number of upsampling filters can be reduced and the transmission filter can be eliminated while improving the effect of upsampling Coefficients and indexes, which can improve the coding performance and improve the effect and performance of image processing. Hereinabove, a method for image processing according to an embodiment of the present invention is described in detail with reference to FIGS. 1 to 3. Hereinafter, a device for image processing according to an embodiment of the present invention will be described in detail with reference to FIGS. 4 to 5. .

Figure 4 shows a schematic block diagram of an apparatus 300 for image processing in accordance with an embodiment of the present invention. As shown in Figure 7, the apparatus 300 includes:

The obtaining unit 310 is configured to determine, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter, so that the first training filter satisfies: according to the first training filter The similarity between the first prediction information determined by the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein The base layer image corresponds to the enhancement layer image; the determining unit 320 is configured to determine, according to the first training filter, an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter Device

For determining from the candidate upsampling filter, the target upsampling filter, and transmitting the target upsampling filter to the encoding unit 330, the candidate upsampling filter including the first training acquired by the acquiring unit 310 Filter

The encoding unit 330 is configured to acquire the upsampling filter from the determining unit, and determine prediction information according to the target upsampling filter and the base layer image block;

And performing, according to the prediction information, performing encoding processing on the target image block to generate a target code stream, where the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and the basic image is The spatial position of the block in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image.

Optionally, the acquiring unit 310 is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter satisfies: The similarity between the second prediction filter determined by the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a second preset condition Wherein the base layer image corresponds to the enhancement layer image;

The determining unit 320 is further configured to determine, according to the first training filter and the second training filter, an alternative upsampling filter, where the candidate upsampling filter comprises the first training filter and the second Training filter.

Optionally, the determining unit 320 is specifically configured to: according to the feature information of the base layer image block, Determining a smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or residual information of the base layer image block;

And configured to determine the target upsampling filter according to the smoothness of the target image block.

Optionally, the determining unit 320 is further configured to determine, according to the first training filter, the second training filter, and the traditional filter, an alternative upsampling filter, where the candidate upsampling filter includes the first a training filter, the second training filter, and the conventional filter;

The encoding unit 330 is specifically configured to perform encoding processing on the target image block according to the prediction information, to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, the first The indication information is used as a basis for acquiring the target upsampling filter when decoding the above-mentioned encoded target image block.

The determining unit 320 is specifically configured to determine a texture degree of the target image block;

And determining the target upsampling filter according to the texture degree of the target image block.

According to the apparatus for image processing according to the embodiment of the present invention, by determining the upsampling filter according to the smoothed region and the non-smoothed region, the number of upsampling filters can be reduced and the transmission filter can be eliminated while improving the effect of upsampling Coefficients and indexes, which can improve the coding performance and improve the effect and performance of image processing.

The apparatus 300 for image processing according to an embodiment of the present invention may correspond to an encoding end in the method of the embodiment of the present invention, and the units in the apparatus 300 for image processing are modules and other operations and/or the above-described operations and/or The functions are respectively implemented in order to implement the corresponding process of the method 100 in FIG. 1 , and are not described herein again.

FIG. 5 shows a schematic block diagram of an apparatus 400 for image processing in accordance with an embodiment of the present invention. As shown in Figure 5, the apparatus 500 includes:

The obtaining unit 510 is configured to determine, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter, so that the first training filter satisfies: according to the first training filter The similarity between the first prediction information determined by the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein The base layer image corresponds to the enhancement layer image; the determining unit 520 is configured to determine, according to the first training filter, an candidate upsampling filter, wherein the candidate upsampling filter comprises the first training filter Device

For determining the target upsampling filter from the alternative upsampling filter, the alternative upsampling filter The wave filter includes the first training filter, and transmits the target upsampling filter to the decoding unit 530, where the candidate upsampling filter includes the first training filter of the acquiring unit acquisition 510;

The decoding unit 530 is configured to acquire the target upsampling filter from the determining unit 520, and determine prediction information according to the target upsampling filter and the base layer image block;

And a decoding process for the target image block according to the prediction information and the residual information obtained from the target code stream, where the target image block is located in the enhancement layer image, and the base layer image block is located in the base layer image. And the spatial position of the basic image block in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image.

Optionally, the acquiring unit 510 is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter satisfies: The similarity between the second prediction filter determined by the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a second preset condition Wherein the base layer image corresponds to the enhancement layer image;

The determining unit 520 is further configured to determine, according to the first training filter and the second training filter, an alternative upsampling filter, where the candidate upsampling filter includes the first training filter and the second Training filter.

Optionally, the determining unit 520 is specifically configured to determine, according to the feature information of the base layer image block, a smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or the base layer. Residual information of the image block;

The determining unit 520 is specifically configured to obtain, from the target code stream, first indication information used to indicate the target upsampling filter,

And determining, according to the first indication information, the target upsampling filter.

The determining unit 520 is specifically configured to determine a texture degree of the target image block;

According to the apparatus for image processing according to the embodiment of the present invention, by determining the upsampling filter according to the smoothed region and the non-smoothed region, the number of upsampling filters can be reduced and the transmission filter can be eliminated while improving the effect of upsampling Coefficients and indexes to improve coding performance Improve the effectiveness and performance of image processing.

The apparatus 400 for image processing according to an embodiment of the present invention may correspond to an encoding end in the method of the embodiment of the present invention, and the units in the apparatus 400 for image processing are modules and the other operations and/or the above-described operations and/or The functions are respectively implemented in order to implement the corresponding processes of the method 200 in FIG. 3, and are not described here.

Fig. 6 shows a schematic block diagram of an encoder 500 for image processing in accordance with an embodiment of the present invention. As shown in FIG. 6, the decoder 500 includes:

Bus 510;

a processor 520 connected to the bus;

a memory 530 connected to the bus;

The processor 520 calls the program stored in the memory 530 through the bus 510 to determine the first training filter according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, to Making the first training filter satisfy: first prediction information determined according to the first training filter and the non-smooth area image block of the base layer image and first original determined according to the non-smooth area image block of the enhancement layer image The similarity between the information satisfies a first preset condition, wherein the base layer image corresponds to the enhancement layer image;

Determining, according to the first training filter, an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter;

Determining a target upsampling filter from the candidate upsampling filter;

Determining prediction information according to the target upsampling filter and the base layer image block;

Obtaining, according to the prediction information, a target image block to generate a target code stream, wherein the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and the basic image block is in The spatial position in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image

Optionally, the processing unit 520 is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter satisfies: The similarity between the second prediction filter determined by the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a second preset condition Wherein the base layer image corresponds to the enhancement layer image;

And determining, according to the first training filter and the second training filter, an alternative upsampling filter The alternate upsampling filter includes the first training filter and the second training filter. Optionally, the processing unit 520 is specifically configured to determine, according to feature information of the base layer image block, a smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or the base layer Residual information of the image block;

Optionally, the processing unit 520 is specifically configured to determine, according to the first training filter, the second training filter, and the traditional filter, an alternative upsampling filter, where the candidate upsampling filter includes the first a training filter, the second training filter, and the conventional filter;

And performing, according to the prediction information, performing encoding processing on the target image block to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, where the first indication information is used to The above-mentioned encoded target image block is decoded as a basis for acquiring the target upsampling filter.

Optionally, the processing unit 520 is specifically configured to determine, according to the first training filter and the traditional filter, an alternative upsampling filter, where the candidate upsampling filter includes the first training filter and the traditional Filter

Used to determine the texture of the target image block;

An encoder for image processing according to an embodiment of the present invention can reduce the number of upsampling filters and reduce the number of upsampling filters while improving the effect of upsampling by determining the upsampling filter according to the smoothed region and the non-smoothed region. The filter coefficients and the index are transmitted, so that the encoding performance can be improved, and the effect and performance of image processing can be improved.

The encoder 500 for image processing according to an embodiment of the present invention may correspond to the encoding end in the method of the embodiment of the present invention, and the units in the encoder 500 for image processing, that is, the module and the other operations described above In order to implement the corresponding process of the method 100 in FIG. 1 , the functions are not described here.

FIG. 7 shows a schematic block diagram of a decoder 600 for image processing in accordance with an embodiment of the present invention. As shown in FIG. 7, the decoder 600 includes:

Bus 610;

a processor 620 connected to the bus;

a memory 630 connected to the bus;

The processor 620 calls the program stored in the memory 630 through the bus 610. Determining, by the non-smooth area image block according to the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter to satisfy the first training filter: according to the first training filter and the The similarity between the first prediction information determined by the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein the basic condition a layer image corresponding to the enhancement layer image;

Determining, from an alternative upsampling filter, a target upsampling filter, the candidate upsampling filter comprising the first training filter;

Decoding the target code stream according to the prediction information and the residual information obtained from the target code stream, to obtain the target image block, where the target image block is located in the enhancement layer image, the base layer image block Located in the base layer image, and the spatial position of the base image block in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image.

Optionally, the processing unit 620 is further configured to determine, according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image, the second training filter, so that the second training filter satisfies: The similarity between the second prediction filter determined by the second training filter and the smooth region image block of the base layer image and the second original information determined according to the smooth region image block of the enhancement layer image satisfies a second preset condition Wherein the base layer image corresponds to the enhancement layer image;

And for determining an alternative upsampling filter according to the first training filter and the second training filter, wherein the candidate upsampling filter comprises the first training filter and the second training filter.

Optionally, the processing unit 620 is specifically configured to determine, according to the feature information of the base layer image block, a smoothness of the target image, where the feature information includes coded block flag information of the base layer image block or the base layer. Residual information of the image block;

Optionally, the processing unit 620 is specifically configured to determine, according to the first training filter, the second training filter, and the traditional filter, an alternative upsampling filter, where the candidate upsampling filter includes the first a training filter, the second training filter, and the conventional filter;

And the first indication information used to indicate the target upsampling filter is obtained from the target code stream, and the target upsampling filter is determined according to the first indication information. . Optionally, the processing unit 620 is specifically configured to determine, according to the first training filter and the traditional filter, an alternative upsampling filter, where the candidate upsampling filter includes the first training filter and the traditional filter;

Used to determine the texture of the target image block;

The decoder for image processing according to an embodiment of the present invention can reduce the number of upsampling filters and eliminate the need to transmit filter coefficients while improving the effect of upsampling by determining the upsampling filter according to the smoothed region and the non-smoothed region. And the index, which can improve the encoding performance and improve the effect and performance of image processing.

The decoder 600 for image processing according to an embodiment of the present invention may correspond to a decoding end in the method of the embodiment of the present invention, and each unit in the decoder 600 for image processing, that is, a module and the other operations described above In order to implement the corresponding process of the method 200 in FIG. 3, the functions are not described here.

It should be noted that, in order to make the encoding end coincide with the target upsampling filter used by the decoding end (encoder, decoder), it is required that the encoding end and the decoding end select the target upsampling filter. In other words, the decoding end processing method corresponding to the described encoding end processing method or the encoding end processing method corresponding to the described decoding end processing method may be determined.

It should be understood that the term "and/or" in this context is merely an association describing the associated object, indicating that there may be three relationships, for example, A and / or B, which may represent: A exists separately, and A and B exist simultaneously There are three cases of B alone. In addition, the character "/" in this article generally indicates that the context object is an "or" relationship.

It should be understood that, in various embodiments of the present invention, the size of the sequence numbers of the above processes does not mean the order of execution, and the order of execution of each process should be determined by its function and internal logic, and should not be taken to the embodiments of the present invention. The implementation process constitutes any limitation.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the various examples described in connection with the embodiments disclosed herein can be implemented in a combination of electronic hardware or computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods for implementing the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the present invention.

It will be apparent to those skilled in the art that for the convenience and cleanness of the description, the specific working processes of the systems, devices and units described above may be referred to the corresponding embodiments in the foregoing method embodiments. The process is not repeated here.

In the several embodiments provided herein, it should be understood that the disclosed systems, devices, and methods may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the unit is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be electrical, mechanical or otherwise.

The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

The functions, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention, which is essential to the prior art or part of the technical solution, may be embodied in the form of a software product stored in a storage medium, including The instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like, which can store program codes. .

The above is only the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the appended claims.

Claims

Rights request

A method for image processing, the method comprising:

Determining, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter to satisfy: the first training filter and the The similarity between the first prediction information determined by the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition, wherein a base layer image corresponding to the enhancement layer image; determining an alternative upsampling filter according to the first training filter, wherein the candidate upsampling filter comprises the first training filter;

Determining a target upsampling filter from the candidate upsampling filter;

Decoding, according to the prediction information, a target image block to generate a target code stream, where the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and A spatial position of the basic image block in the base layer image corresponds to a spatial position of the target image block in the enhancement layer image.

The method according to claim 1, wherein the method further comprises: determining a second training filter according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image So that the second training filter satisfies: second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the smooth region image block according to the enhancement layer image The similarity between the second original information satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image;

Determining the candidate upsampling filter according to the first training filter, further comprising: determining, according to the first training filter and the second training filter, an alternative upsampling filter, where The alternative upsampling filter includes the first training filter and the second training filter.

The method according to claim 2, wherein the determining the target upsampling filter from the candidate upsampling filter comprises:

Determining a smoothness of the target image according to the feature information of the base layer image block, where the feature information includes coded block flag information of the base layer image block or residual information of the base layer image block ; The target upsampling filter is determined according to the smoothness of the target image block.

The method according to claim 2, wherein the determining the candidate upsampling filter according to the first training filter and the second training filter further comprises:

Determining an alternative upsampling filter according to the first training filter, the second training filter, and a conventional filter, wherein the candidate upsampling filter includes the first training filter, a second training filter and the conventional filter;

And performing the encoding process on the target image block to generate the target code stream according to the prediction information, including:

Decoding, according to the prediction information, the target image block to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter, the first indication information It is used as a basis for acquiring the target upsampling filter when decoding the above-mentioned encoded target image block.

The method according to claim 1, wherein the determining the candidate upsampling filter according to the first training filter further comprises:

Determining an alternative upsampling filter according to the first training filter and the conventional filter, wherein the candidate upsampling filter comprises the first training filter and the conventional filter; and the determining The target upsampling filter includes:

Determining a texture degree of the target image block;

6. A method for image processing, the method comprising:

Determining, from an alternative upsampling filter, a target upsampling filter, the alternative upsampling filter comprising the first training filter;

And the target code stream is input according to the prediction information and residual information obtained from a target code stream a row decoding process to obtain the target image block, wherein the target image block is located in the enhancement layer image, the base layer image block is located in the base layer image, and the basic image block is in the The spatial position in the base layer image corresponds to the spatial position of the target image block in the enhancement layer image.

The method according to claim 6, wherein the method further comprises: determining a second training filter according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image So that the second training filter satisfies: second prediction information determined according to the second training filter and the smooth region image block of the base layer image and the smooth region image block according to the enhancement layer image The similarity between the second original information satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image;

The method according to claim 7, wherein the determining the target upsampling filter from the candidate upsampling filter comprises:

Determining a smoothness of the target image according to the feature information of the base layer image block, where the feature information includes coded block flag information of the base layer image block or residual information of the base layer image block ;

The method according to claim 8, wherein the determining the candidate upsampling filter according to the first training filter and the second training filter further comprises:

Determining the target upsampling filter from the alternative upsampling filter, comprising:

Obtaining, from the target code stream, first indication information for indicating the target upsampling filter,

And determining the target upsampling filter according to the first indication information.

10. The method according to claim 6, wherein said according to said first training A filter, determining an alternative upsampling filter, further comprising:

Determining a texture degree of the target image block;

11. An apparatus for image processing, the apparatus comprising:

And an obtaining unit, configured to determine, according to the non-smooth area image block of the enhancement layer image and the non-smooth area image block of the base layer image, the first training filter, so that the first training filter satisfies: according to the first training The similarity between the first prediction information determined by the filter and the non-smooth area image block of the base layer image and the first original information determined according to the non-smooth area image block of the enhancement layer image satisfies a first preset condition Wherein the base layer image corresponds to the enhancement layer image;

a determining unit, determining, according to the first training filter, an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter;

For determining a target upsampling filter from the candidate upsampling filter;

a coding unit, configured to acquire the upsampling filter from the determining unit, and determine prediction information according to the target upsampling filter and the base layer image block;

And configured to perform encoding processing on the target image block according to the prediction information, to generate a target code stream, where the target image block is located in the enhancement layer image, and the base layer image block is located in the base layer image. And a spatial position of the basic image block in the base layer image corresponds to a spatial position of the target image block in the enhancement layer image.

The device according to claim 11, wherein the acquiring unit is further configured to determine a second training filter according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image So that the second training filter satisfies: second prediction information determined according to the second training filter and the smooth region image block of the base layer image and a smooth region image block according to the enhancement layer image The determined similarity between the second original information satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit is further configured to filter according to the first training And the second training filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the second training filter.

The device according to claim 12, wherein the determining unit is configured to determine a smoothness of the target image according to feature information of the base layer image block, where the feature information includes Encoding block flag information of a base layer image block or residual information for the base layer image block;

The device according to claim 12, wherein the determining unit is further configured to determine an alternative upsampling filter according to the first training filter, the second training filter, and a conventional filter. The candidate upsampling filter includes the first training filter, the second training filter, and the conventional filter;

The coding unit is specifically configured to perform coding processing on the target image block according to the prediction information, to generate a target code stream, where the target code stream includes first indication information for indicating the target upsampling filter. The first indication information is used as a basis for acquiring the target upsampling filter when decoding the encoded target image block.

The device according to claim 11, wherein the determining unit is further configured to determine an alternative upsampling filter according to the first training filter and the traditional filter, wherein the candidate is a sampling filter comprising the first training filter and the conventional filter;

Specifically for determining a texture degree of the target image block;

16. An apparatus for image processing, the apparatus comprising:

a determining unit, configured to determine, according to the first training filter, an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter;

For determining from the candidate upsampling filter, the target upsampling filter, the candidate upsampling filter comprising the first training filter; the target upsampling filter and the base layer image block, determining prediction information ; And performing decoding processing on the target code stream to obtain the target image block according to the prediction information and residual information acquired from a target code stream, where the target image block is located in the enhancement layer image The base layer image block is located in the base layer image, and a spatial position of the basic image block in the base layer image corresponds to a spatial position of the target image block in the enhancement layer image .

The device according to claim 16, wherein the acquiring unit is further configured to determine a second training filter according to the smooth region image block of the enhancement layer image and the smooth region image block of the base layer image So that the second training filter satisfies: second prediction information determined according to the second training filter and the smooth region image block of the base layer image and a smooth region image block according to the enhancement layer image The determined similarity between the second original information satisfies a second preset condition, wherein the base layer image corresponds to the enhancement layer image; and the determining unit is further configured to filter according to the first training And the second training filter, determining an alternative upsampling filter, wherein the candidate upsampling filter comprises the first training filter and the second training filter.

The device according to claim 17, wherein the determining unit is configured to determine a smoothness of the target image according to the feature information of the base layer image block, where the feature information includes Encoding block flag information of a base layer image block or residual information for the base layer image block;

The device according to claim 17, wherein the determining unit is further configured to determine an alternative upsampling filter according to the first training filter, the second training filter, and a conventional filter. The candidate upsampling filter includes the first training filter, the second training filter, and the conventional filter;

And acquiring, from the target code stream, first indication information used to indicate the target upsampling filter,

The device according to claim 16, wherein the determining unit is further configured to determine an alternative upsampling filter according to the first training filter and the traditional filter, wherein the candidate is a sampling filter comprising the first training filter and the conventional filter;

Used to determine the texture degree of the target image block;