KR100466595B1 - Shape Information Encoding Method of Object Image - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to a method of encoding shape information of an object image.

2. The technical problem to be solved by the invention

According to the present invention, when content-based image coding is applied, shape information loss coding is applied and there is no original image shape information and only reconstructed shape information exists, by applying a padding process to a block containing only reconstructed shape information, garbage information is encoded as it is. The purpose of the present invention is to provide a method of encoding shape information of an object image, which can prevent encoding, and improve encoding efficiency and image quality.

3. Summary of Solution to Invention

According to the present invention, a shape information encoding method for processing an image applied through an image input apparatus into a target image and a background image having arbitrary shape information is processed. If only the reconstructed shape information exists, the luminance component in the block in which only the reconstructed shape information exists when intra texture coding and / or inter texture coding of an image signal and / or an error signal is performed. (Y) and the two color components (Cb, Cr) are separated and encoded by padding with arbitrary values.

4. Important uses of the invention

The present invention is used in MPEG-4 system and the like.

Description

Shape information encoding method of object image

The present invention relates to a shape information encoding method for processing a video applied through an image input apparatus into a target image and a background image having arbitrary shape information. In the case of intra-encoding, the luminance information and the two color information are padded with 128 for blocks that do not have the original image shape information but only the restored shape information. Therefore, to improve the coding efficiency and the image quality.

As is well known, recent image processing techniques break away from a method of compressing and encoding an image corresponding to one frame as a whole, dividing the image into predetermined unit blocks having arbitrary shape information, and then compressing and encoding each of them. In the direction of transmission.

That is, the applied image is separated into each object image and the background image, and only the change of the object image is transmitted to achieve compression efficiency and encoding efficiency, and an international standard for this is being prepared.

For example, WG11 under ISO / IEC, the world's standardization organization, encodes an object with arbitrary shape information, unlike MPEG-2 (International Standard for Moving Picture Experts Group) -1 and MPEG-2. MPEG-4, which is a standardization work in progress, is being carried out, and MPEG-4, which is being standardized, is a unit block having arbitrary shape information and is based on the concept of VOP (Video Object Plane).

Here, the VOP is defined as a quadrangle that divides the applied image into a background image and each object image and includes the separated background image and the object image. In MPEG-4, the VOP is defined as a predetermined object or a predetermined region in the image. If there is a region of the object to be made up, it is a skeleton to separate the image of the object into respective VOPs and to encode the separated VOPs respectively.

The VOP has the advantage of freely synthesizing or decomposing a natural image or an artificial image as a unit of an object image, and is the basis for processing an object image in the field of computer graphics and multimedia.

For example, applications currently expected to apply MPEG-4 technology include intercommunication such as Internet multimedia (IMM), interactive video games (IVG), video conferencing, and video telephony. Network database services (IPB: Interpersonal Communications), Interactive Storage Media (ISM), Multimedia Mailing (MMM), Wireless Multimedia (WMM), ATM networks, etc. Database Service), Remote Emergency Systems (REs), and Remote Video Surveillance (RVS).

In order to support these existing or future applications, a video encoding technique is required that allows users to communicate only with the desired objects in the image, access them to find and read them, and edit them to cut and paste them. . MPEG-4, a new image and audio coding technology underway with global standardization, aims to meet this need.

1 illustrates a structure of an image encoder of MPEG-4, which has a structure different from that of H.261, H.263, and MPEG-2, which are the world standardizations of image encoding. In particular, the introduction of the concept of the shape coder (Shape Coder) and VOP shows the most noticeable difference. A VOP refers to an object at a point in time on an arbitrary shape of a content that can be accessed and edited by a user, and must be encoded for each VOP to support content-based functionality.

2 shows an example of a VOP having an arbitrary shape. The VOP is reconstructed into macroblocks of 16x16 pixels for picture coding. At this time, three types of macroblocks exist in the reconstructed VOP, as shown in FIG. 3. In this case, macroblocks in an object composed of only information in the object, macroblocks outside the object having no information in the object, and objects It is an object boundary macroblock that partially has information in it. In addition, the arrangement of the luminance block and the color block in each macro block is as shown in FIG. Intra-object macroblocks consist only of intra-object blocks, and out-of-object macroblocks consist only of out-of-object blocks.

In the object boundary macroblock, there can be three kinds of blocks composed of 8x8 pixels having the same characteristics as the macroblock type as shown in FIG. 5. This is it. In this case, in order to encode the image information and the prediction error information of the object boundary block, the pixel values of the pixels in the out-of-object region must be determined first.

In MPEG-4 Verification Model (VM) 7.0, Low-Pass Extrapolation Technique (LPE) using intra-object image information is performed for intra-image encoding, and zero-padding technique for encoding prediction error information during intra-image encoding. Perform Zero Padding Technique.

The LPE technique is a technique of encoding an information of an object outside an object by filtering a low frequency component and then encoding the information of an object outside an object with a zero value. Technology.

In other words, VM7.0 uses LPE padding for intra-image encoding using shape information of the original image on a block basis, and zero padding for inter-image encoding, and then blocks the block using DCT (Discrete Cosine Transform). However, when loss coding the shape information, there may be a block in which there is no shape information of the original image and only shape information of the reconstructed image. Since there is no original image shape information, the padding technique of VM7.0 is applied to the block. This can't be applied.

Therefore, since garbage information that is not related to image information is encoded and decoded, both coding efficiency and image quality deteriorate.

That is, the encoding of the object boundary block of MPEG-4 VM7.0 (ISO / IEC N1642, "Text of MPEG-4 Video VM7.0") applies DCT technology using LPE technology, zero padding technology, etc. While the block is DCT encoded without padding, the object boundary block must first be padded with an arbitrary value using the shape information of the original image for DCT encoding.

However, in the case of loss coding the shape information, a block having only the shape information of the original image without reconstruction may be generated. Since the padding process cannot be applied, the block information must be encoded as it is. As a result, there is a problem that the coding efficiency and image quality is reduced.

The present invention has been proposed to solve the above problems, and when content-based image coding is applied, shape information loss coding is applied so that there is no original image shape information and only reconstructed shape information exists, padding is provided on the block having only the reconstructed shape information. It is an object of the present invention to provide a method of encoding shape information of an object image, which can prevent garbage information from being encoded as is, thereby improving encoding efficiency and image quality.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. Also, it will be readily appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

Apparatus of the present invention for achieving the above object, in the shape information loss method in the shape information encoding method for processing by separating the image applied through the image input device into each object image and the background image having arbitrary shape information, If the original image shape information is not applied and only reconstructed shape information exists, the reconstructed shape is obtained when intra texture coding and / or inter texture coding are performed on an image signal and / or an error signal. The luminance component (Y) and the two color components (Cb and Cr) in a block in which only information is provided are distinguished from each other and padded with an arbitrary value.

When encoding the shape information, if the lossy coding method is applied, the restored shape information and the shape information of the original image are inconsistent with each other as shown in FIG. 6, and at this time, only the shape information of the reconstructed image may occur. This technique is applied to blocks with recovered shape information only.

Accordingly, in the present invention, when intra texture coding is applied to a block having only reconstructed shape information as it is, the luminance component Y and the two color components of the entire block are applied. In the case where DCT coding is performed by padding Cb and Cr to 128, 128, and 128, and inter texture coding for encoding prediction error information is applied, the luminance component Y of the block is determined. With 0, two color components are padded with 128 and DCT encoded.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

According to the present invention, when intra texture coding is applied to a block having only reconstructed shape information as it is, the luminance component Y and the two color components Cb, In the case where DC texture is performed by padding each of the two bits to 128, and inter texture coding for encoding prediction error information is applied, the luminance component (Y) of the block is zero, and two color components are used. (Cb, Cr) are each padded with 128 to perform DCT coding (first embodiment).

In the case where there is a macroblock without original image shape information and only reconstructed shape information, a method of applying a macroblock unit instead of a padding and encoding method proposed in the present invention, that is, reconstructing without original image shape information For macroblocks with only shape information, the luminance component (Y) and both color components (Cb, Cr) are padded at 128 during intra texture coding, and the luminance component at inter texture coding. DCT coding is performed by padding (Y) to 0 and two color components (Cb and Cr) to 128, respectively (second embodiment).

In addition, when the SADCT (Shape-Adaptive Discrete Cosine Transform) technology is used instead of the DCT technology, the method of applying the block padding and the macroblock padding proposed in the 'shape information encoding method of the object image' is applied. Can be applied.

For reference, the color is represented using a method called YCbCr. In general, RGB is a method, which is a method of expressing a color by mixing colors of red light (R), green light (G), and blue light (B), which are three elements of light. However, YCbCr is a method of expressing a color using luminance Y, which is the intensity of color, color difference information Cb with blue, and color difference information Cr with red.

Meanwhile, FIG. 1 illustrates the structure of an MPEG-4 VOP encoder, and illustrates a structure of a VOP encoder considered in MPEG-4, which is a coding standard for next-generation video and audio information. In addition, FIG. 2 shows a VOP, wherein the VOP is composed of an entire image or an image composed only of meaningful objects forming the image.

3 illustrates the types of macroblocks in a VOP image. That is, the VOP is reconstructed in units of macroblocks for encoding, and FIG. In addition, each macroblock of FIG. 3 is composed of an out-of-object macroblock, an in-object macroblock, and an object boundary macroblock according to the presence or absence of image information.

4 shows the arrangement of blocks in the macroblock, with the upper left side as block 1, the upper right side as block 2, the lower left side as block 3, the lower right side as block 4, the color information Cb as block 5, and color. The information Cr is defined as block 6, respectively.

5 shows the types of blocks in a VOP picture. That is, out-of-object macroblocks of the macroblocks in the reconstructed VOP image as shown in FIG. 3 are composed of only out-of-object blocks, and intra-object macroblocks are composed of only intra-object blocks. However, the object boundary macroblock is composed of an intra-object block, an object boundary block, or a block out of the object.

FIG. 6 illustrates an example of a block in which no original image shape information is present and only reconstructed shape information exists when loss coding of shape information is applied, and FIG. 7 shows a process of encoding image information of the existing technology.

8 illustrates a shape information encoding method according to the present invention.

As shown in FIG. 8, the encoding of an object boundary block of MPEG-4 VM7.0 applies a DCT technique using an LPE technique, a zero padding technique, and the like. In this case, a block within an object may be DCT encoded without padding. The boundary block should be padded to an arbitrary value using the shape information of the original image for DCT encoding. However, in the case of loss coding the shape information, a block having only reconstructed shape information without the shape information of the original image may be generated. In the present invention, the shape information loss encoding is applied when the content-based image encoding is applied. If there is no image shape information and only reconstructed shape information exists, the luminance component (Y) and the two color components (Cb, Cr) in the block in which only the reconstructed shape information exist when encoding the image signal and the error signal, respectively Separated and padded with arbitrary values.

For example, during intra texture coding, DCT is padded with the same luminance component (Y) and two color components (Cb, Cr) in a blick in which there is no original image shape information and only reconstructed shape information. Encode For example, the luminance component Y and the two color components Cb and Cr are padded with 128 to be DCT encoded.

In inter texture coding, the luminance component (Y) and the two color components (Cb, Cr) in a block in which there is no original image shape information and only reconstructed shape information are padded with different values to perform DCT coding. do. For example, DCT coding is performed by padding both luminance components Y with 0 and padding both color components Cb and Cr with 128.

As another example, when intra texture coding, the luminance component (Y) and the two color components (Cb, Cr) in a macroblock in which there is no original image shape information but only restored shape information are padded with the same value. DCT coded. For example, the luminance component Y and the two color components Cb and Cr are padded with 128 to be DCT encoded.

In the inter texture coding, the luminance component (Y) and the two color components (Cb, Cr) in the macroblock in which the original image shape information is restored and only the restored shape information exist are padded with different values. Encode For example, DCT coding is performed by padding both luminance components Y with 0 and padding both color components Cb and Cr with 128.

As described above, the method of the present invention may be implemented as a program and stored in a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.) in a computer-readable form. Since this process can be easily implemented by those skilled in the art will not be described in more detail.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

In the present invention as described above, the padding process is applied to a block having only the reconstructed shape information without modifying the shape information encoding technique of MPEG VM7.0, that is, the reconstruction is performed by the shape information encoding technique during VOP-based image encoding. When a block having only the shape information is generated, DCT coding is performed by padding both the luminance component (Y) and the two color components (Cb, Cr) to 128 during intra texture coding for the entire block. In inter texture coding, the luminance component (Y) is 0 and the two color components (Cb, Cr) are padded with 128 to DCT encoding, thereby preventing garbage information from being encoded as it is, thereby improving coding efficiency and image quality. It can increase the effect.

1 is a block diagram showing the structure of an MPEG-4 VOP encoder;

2 is an explanatory diagram showing an example of a VOP;

3 is an explanatory diagram showing types of macroblocks in a VOP image;

4 is an explanatory diagram showing an arrangement of blocks in a macroblock;

5 is an explanatory diagram showing types of blocks in a VOP video;

6 is an explanatory diagram showing an example of a block in which no original image shape information exists and only reconstructed shape information exists when loss encoding of shape information is applied;

7 is an explanatory diagram showing a video information encoding process according to the prior art;

8 is an exemplary diagram illustrating a shape information encoding method according to the present invention.

Claims (9)

In the shape information encoding method for processing by separating the image applied through the image input device into each of the object image and the background image having arbitrary shape information, When the shape information loss encoding is applied and there is no original image shape information and only reconstructed shape information exists, when intra texture coding and / or inter texture coding are performed on an image signal and / or an error signal. And encoding the luminance component (Y) and the two color components (Cb and Cr) in the block in which only the reconstructed shape information exists, and padding and encoding them by arbitrary values. The method of claim 1, In the intra texture coding, An object image, characterized in that a discrete cosine transform (DCT) encoding is performed by padding both luminance components (Y) and two color components (Cb, Cr) in a block in which there is no original image shape information and only reconstructed shape information. Shape information encoding method of a. The method of claim 2, In the intra texture coding, 12. A method of encoding shape information of an object image, characterized in that the luminance component (Y) and the two color components (Cb, Cr) in a block in which there is no original image shape information and only reconstructed shape information are padded by 128 are DCT encoded. The method of claim 1, In the inter texture coding, An object image, characterized in that a discrete cosine transform (DCT) encoding is performed by padding the luminance component (Y) and the two color components (Cb, Cr) in a block in which there is no original image shape information and only reconstructed shape information. Shape information encoding method of a. The method of claim 4, wherein In the inter texture coding, An object image, characterized in that the luminance component (Y) in a block in which there is no original image shape information and only reconstructed shape information is padded with 0, and both color components (Cb, Cr) are padded with 128 to be DCT encoded. Shape information encoding method of a. The method of claim 1, In the intra texture coding, An object characterized in that a discrete cosine transform (DCT) coding is performed by padding the luminance component (Y) and the two color components (Cb, Cr) in a macroblock in which the original image shape information is restored and only the restored shape information exists. Shape information encoding method of image. The method of claim 6, In the intra texture coding, Shape information encoding method of the target image, characterized in that the luminance component (Y) and the two color components (Cb, Cr) in the macroblock in which there is no original image shape information and only reconstructed shape information are padded with 128 to be DCT encoded. . The method of claim 1, In the inter texture coding, An object characterized in that a discrete cosine transform (DCT) encoding is performed by padding the luminance component (Y) and the two color components (Cb, Cr) in a macroblock in which only the restored shape information exists without the original image information. Shape information encoding method of image. The method of claim 8, In the inter texture coding, An object characterized in that all of the luminance components (Y) in a macroblock in which there is no original image shape information but only reconstructed shape information are padded with 0, and both color components (Cb, Cr) are padded with 128 to be DCT encoded. Shape information encoding method of image.

Images