JPH02226486A - Picture encoding/synthesizing system - Google Patents

Abstract

PURPOSE:To prevent an objective picture from being hidden by a background picture when the pictures are synthesized by segmenting the objective picture from the component picture, and discriminating the objective picture from the background picture in the component picture. CONSTITUTION:In three component pictures P1, P2 and P3 surrounded by dot lines, respective hatched portions are objective pictures O1, O2 and O3, and the other parts are background pictures b1, b2 and b3. In the component pictures P1 to P3, m-bit codes are allocated to respective picture elements as area information to discriminate the backgrounds b1 to b3 from the objects O1 to O3, and by compressing only the part indicated to be the object by the area information, an information quantity and a calculating time can be reduced. Further the area information can discriminate the objective pictures O1 to O3 in the component pictures P1 to P3 from the background picture parts b1 to b3. Thus even when the background pictures of the other component picture are overlapped on the objective pictures O1 to O3 of the component pictures P1 to P3, the data of the objective pictures O1 to O3 can be prevented from being erased.

Description

[Detailed Description of the Invention] <Industrial FLL Field> The present invention encodes and saves a component image and an entire background image,
This invention relates to an image encoding and combining method for combining images.

<Conventional technology> Conventionally, image synthesis has been performed as follows.

That is, the entire background image B shown in FIG.
A rectangular component image PI shown in FIG. 1(b).

A composite image A is obtained by superimposing P, , P, and l as shown in FIG. 11(c).

The following methods are available for encoding, storing, and restoring this composite image A.

(2) As shown in FIG. 12(a), a method in which the entire composite image A is regarded as one image, compression processing is performed, encoded data is recorded and saved, and this encoded data is restored and displayed.

■ As shown in FIG. 12(b), in addition to the entire background image B, the component image P2. Perform compression processing for each Pt and Ps,
When recording encoded data and composing images, the encoded data is restored and each component image P=, P2 . A method of replacing data at a predetermined position of the entire background image B with the data of P3 and displaying it.

<Problems to be Solved by the Invention> However, the above-mentioned conventional image coding and synthesis method has the following problems.

That is, in the method (■), in order to save the entire composite image A, data sharing of the entire background image B and component image P are performed.
1. P! , P3 cannot be shared and must be saved for each overall composite image A, resulting in a disadvantage that the amount of data to be saved becomes enormous. Another problem is that the component images P, , P, , P3 cannot be handled individually and moved freely.

In addition, in both methods ■ and ■, when the part images P and A part images Pt are displayed in a superimposed manner as shown in FIG. 13, the target images 0. ．． 0. and the background image b=,b in the part images P, ,Pr
Since the object image aO1 cannot be distinguished from the background image bt, a problem arises in that the target image aO1, which is below the background image bt, is hidden, resulting in an invisible portion. Alternatively, as shown in FIG. 14, if there are holes H+ and Ht inside the target image O2, the entire background image B should originally be visible when combined)r,
, H, were sometimes hidden by the background image S of the component image P, and the entire background image B could no longer be seen. Furthermore, in FIG. 12(b), even if the image data is compressed and encoded using the method (3) and stored, the component image P1. Pt,
P3 has target image 0. consisting of a plant, a frame, and a chair. ．． 0.
．． Since the background images S, b, and b9 are included in addition to 0°, there is a problem that there is a lot of useless information even if it is compressed, and the calculation time is long.

Therefore, an object of the present invention is to separate the entire background image and the part image, or between the part images and the part image, so that the target image in the part image is not hidden by the background image, and the hole part in the part image is completely covered. It is possible to perform image synthesis so that the background image is represented, it is possible to handle each part image separately, there is no positional shift or deterioration of image quality, and it is possible to reduce the amount of data to be compressed and encoded and the amount of calculation. The purpose of this invention is to provide an image encoding and synthesis method that can perform the following steps.

<Means for Solving the Problems> In order to achieve the above object, the image encoding synthesis method of the present invention cuts out a target image from a component image, and converts each pixel in the component image into the target image.

Create region information to determine whether the image is included in the background image of the component image, and the color information of each pixel of the entire background image and the color information of each pixel of the target image in the component image are created. While reversibly compressing and storing, the region information of each pixel of the component image is reversibly compressed and stored, the irreversibly compressed color information is decoded, and the reversibly compressed region information is decoded, The target image in the component image and the entire background image are synthesized based on the color information of each pixel of the target image in the component image, the area information, and the color information of each pixel of the entire background image. There is.

Further, it is preferable that the area information includes edge information indicating a boundary between the target image and the background image in the component image, and color information of the boundary is created based on the edge information. desirable.

Furthermore, the color information of the pixels in the boundary is created by interpolation based on the edge information from the color information of the pixels of the target image near the boundary and the color information of the pixels of the entire background image near the boundary. It is desirable to do so.

According to the present invention, an m-bit code is assigned to each pixel of a component image as area information for distinguishing the background from the target, and only the part where the area information indicates that it is the target is compressed. By love? ! The amount and calculation time can be reduced. In addition, since it is possible to distinguish between the target image part of a part image and the non-target image part of the part image based on the area information, even if the background image of another part image is superimposed on the target image of one part image, the data of the target image can be prevented from disappearing. Additionally, the entire background image can be displayed in the hole portion of the target image. To compress region information, a reversible compression method is used because the visual characteristics of humans have high spatial resolution and are extremely sensitive to positional shifts. On the other hand, color information is irreversibly compressed because the visual sense is not so sensitive. Furthermore, in order to prevent a color mismatch between the target image in the partial image and the entire background image from occurring as shown in Figure 15 by compositing two or more images, the region information is added to the target image. By adding edge information and applying a color to the edges that is a combination of the color information of the entire background image and the color information of the target image of the component image, inconsistencies at the boundaries are made less noticeable. .

In this way, the synthesized image created by the above-mentioned image encoding synthesis method has a small amount of information when stored and a small amount of calculation time for encoding and decoding, and can be accurately and freely superimposed on each component image. Furthermore, since there is no mismatch at the boundary between the target image of the component image and the entire background image, a good-looking composite image can be obtained.

<Embodiment> FIG. 1 shows a schematic diagram of a system configuration for realizing the present invention. 1.2 is image data such as a manuscript, painting, photograph, etc. that will be processed from now on. 3 is an image input device such as a color scanner that takes in the image data 1.2. Reference numeral 4 denotes a cutout processing unit that creates area information by performing processing to distinguish between a background image and a target image when a component image is input.

Reference numeral 5 denotes an irreversible compression unit that performs compression processing of color information of the target image in the entire color background image and component images, and numeral 6 represents a reversible compression unit that performs compression processing of area information of the component image.

7 is a hard disk (HD) and a floppy disk (FD) for recording and saving data that has been compressed.

This is an image recording and reproducing device such as magnetic tape (MT). The above is the encoding processing section. The decoding processing section will be described below. Reference numeral 8 denotes an irreversible decoding processing unit that restores the color information of the entire color background image and the target image in the component image out of the data read out from the recording/reproducing device 7, and 9 a reversible decoding processing unit that restores the area information of the component image. 10 is a synthesis processing unit that synthesizes the entire color background image and the component image;
Reference numeral 11 denotes an image display device that displays images synthesized by the synthesis processing section IO.

Next, the processing of each part will be described in detail.

When the human image is designated as a component image, the cutout processing unit 4 distinguishes the target image and background image areas. Three parts images P surrounded by dotted lines as shown in FIG. 2 1. The shaded areas for P t and P 3 are target images 0. ．． 0! , 03, the others are background images b+,
bt, b3. From now on, I will talk about the process of treating the plants. At this time, as shown in Figure 3, the area information is 0 for the background image and 1 for the target image.
Assign. This processing is performed by the user interactively determining 0 and 1, and as shown in FIG. 4, area information is stored in the lower m bits of the image data. Therefore, if the image data is Q bits/pixel, the remaining (σ-m) bits are used to store color information. Therefore, the information of the lower m bits of the image data read from the image human input device 3 will be lost. However, when fishing on a boat, human vision hardly feels any deterioration if the information in the lower 2 to 3 bits of each 8-bit R, G, and B pixel disappears, so if the value of m is small, the effect on vision is small. rare. The region information created by the cutout processing section 4 is sent to the reversible compression section 6. Since human visual characteristics have a high spatial resolution, they are very sensitive to deviations in the position (that is, edges) of the line 9 plane, and therefore, the region information must be restored with high accuracy.

Therefore, a reversible method such as run-renoguess coding is used to compress the region information. Color information such as the target image and the entire color background image in the component image is sent to the irreversible compression unit 5. Human visual characteristics are not very sensitive to color information, so for example, G B T C
(Generalized Block Trunk
ion Coding), AD CT (Adapt
ive Discrete Co51ne Trans
form), PCS (ProgressiveCo
An irreversible method such as ding Scheme) is used.

The data compressed by the irreversible compression section 5 and the reversible compression section 6 is written to the recording/reproducing device 7 and stored.

Next, we will discuss decoding. The color information of the image data read from the recording/reproducing device 7 is decoded by the irreversible decoding section 8, and the area information of the component image is decoded by the reversible decoding section 9. The decoded entire color background image and component image are combined in a combination processing section 10. Here, the area information of the component image is referred to, and in the case of 0 (background), the color information of the entire color background image is stored in the image memory of the composition processing unit IO. On the other hand, when the area information is 1 (target image), the color information of the component image is stored in the image memory. The image synthesis is completed by displaying the color information stored in the image memory on the image display device 11 as shown in FIG. If the area information is O, the entire color background image is displayed as is, so it is possible to output a natural-looking image even when part images overlap or there are holes in the part image. It is.

However, in the area information, there are two background (0) and target (1)
Since there is only one distinction, color mismatch often occurs at the boundary between these two areas, as shown in FIG.

Therefore, an edge portion is further provided in the region information, and in this boundary portion, a neutral color between the background and the object is obtained by interpolation, thereby suppressing edge mismatch. For example, the area information is the background (0
), edge (12), and target (3). The difference between edges (1, 2) is that the intermediate color (1) is close to the background and the intermediate color (2) is close to the object.

This will be explained below using a more specific example.

The three color components R and G are used as the entire color background image data (Figure 11(a)) and the component image data (Figure 1I(b)).
, B values Cr, Cg, and Cb are taken in from the image human input device. C1 (i=r, g, b) is 8 bits and takes a value from 0 to 2'-1. The component image undergoes the following processing in the cutout processing section. When the user performs the cutting process interactively, as shown in Figure 6, the region information is set to 0 for the background image, the region information is 3 for the target image, and the location for the edges. Area information by! or 2, and area information A 1 (i = O ~ 3
, Ai is 2 bits) are created. Next, image data C1 (
The lower two bits of i=r, g, b) are replaced with area information Ai, and the upper six bits are used as color information Ci°. In other words, the component image data X i (X i: 8 bits) is expressed by the following equation.

Ci' -〇 i & F C (IIXc i' is the color information of the component image, (H) is a hexadecimal number.) Xi = C4' + Ai (Ai is area information) The entire color background image data is a normal irreversible code (e.g. GBT
C, ADCT, VQ, PO2, etc.) and is written to the recording/reproducing device 7'. However, at this time, it is necessary to set a discrimination flag indicating whether this image is an entire color background image or a component image. The component image Xi is separated into the upper 6 bits 1-Ci' and the lower 2 bits AI, the color information Ci is sent to the irreversible compression section 5, and only the pixels whose area information is 3 (target image) are encoded. In other words, no encoding processing is performed on the background and edges.

Since the color information CI' of the pixel whose area information is 3 is 6 bits, it is shifted to the left by 2 bits to become 8 bits Ci'', and subjected to the same irreversible compression encoding process as the entire color background image data.

Ci”=Ci°×22 The area information Ai of the component image is reversibly encoded in the reversible compression unit 6. An example is shown below. The data in FIG. and 3, followed by the same code.The code length n is n = [logtN + 1] - [1og29 + if the width of the part image is N pixels.
1]=4. In addition,[
] is a Gaussian symbol. Also, the edge (1, 2) is (1,
2) The code of 2 bits is sent as long as 1 or 2 continues. FIG. 7 shows the result of encoding the area information in FIG. 6.

The above encoded data is written to the recording/reproducing device 7.

From now on, this entire color background image and component image can be displayed not only in combination with these two images, but also in combination with other data previously stored in the recording/reproducing device 7. In the decoding process of the entire color background image, the flagged data of the color background image is read out from the recording/reproducing device 7 and sent to the irreversible decoding processing section 8 . Here, the entire color background image data is reproduced from the encoded data. This image data is stored in the image memory of the temporary combining section 10.

In the component image decoding process, the recording/reproducing device 7 successively outputs flagged component image data, and the reversible decoding processing section 9 decodes the area information shown in FIG. 6(a).

While referring to this area information, the irreversible decoding unit 8 decodes the color information of the target image of the component image. In this way, the color information of the component image is decoded only for pixels whose area information is 3 (target image), and the result is as shown in FIG. 8(a). This is stored in an image memory separate from the image memory that stores the entire color background image. When compositing these two images, the user selects the starting point (XO, yo) of the part image (
Figure 8(a)) can be placed at any position (X
S, ys) (FIG. 8(b)). The composition processing unit 10 performs a process of replacing the pixels Ci of the entire color background image corresponding to the area information 3 (target image) of the component image with the color information Si of the component image. Area information is θ~2
No processing is performed when . FIG. 9(a) shows an enlarged view of FIG. 8(b) when the above processing is performed. Next, process the edges. The area information of the part image is for pixels l and 2 (
Find the pixel of the color entire background image (see FIG. 9(b)) that corresponds to the pixel (see FIG. 6(a)), and find the target image of the pixel of the entire background image and the part image whose area information closest to that pixel is 0. The pixels inside are found and their color information values are set as representative values Hi and Ti of the background and object. If there are a plurality of pixels closest to the pixel of interest, the average value of each pixel is taken as the representative value. This representative value Hi.

The edge pixel value Ei is determined from Ti using the following formula.

Here, Z is edge area information (l or 2).

m-1 As an example, the value of edge A in FIG. 9(b) is determined as follows. Since the background pixel closest to A is C, H8-2
Since the target pixel closest to 0 and A is D, TA=I
OO1 therefore becomes. The results thus obtained are shown in FIG. 1O.

This processing makes it possible to suppress the occurrence of color mismatch between the object and the background at the edges, and it is possible to display an attractive image on the image display device 11.

<Effects of the Invention> As is clear from the above, according to the present invention, the target image is cut out from the part image and the target image and the background image are distinguished in the part image. Images are not hidden behind background images, and
The entire background image can be displayed in the hole part in the target image.

Further, according to the present invention, the color information of each pixel of the entire background image and the color information of each pixel of the target image in the component image are irreversibly compressed and stored, while the area information of each pixel of the component image is stored. Since the data is reversibly compressed and stored, the amount of data m and calculations to be compressed and encoded and decoded can be reduced, and high-quality image synthesis can be performed with almost no positional shift or deterioration in image quality.

Further, according to the present invention, the area information includes edge information indicating the boundary between the target image and the background image in the component image, and the color information of the pixels at the boundary is applied to the target image near the boundary. Since the color information is created by interpolation based on the edge information from the color information of the pixels in the border and the color information of the pixels of the entire background image near the border, there will be no color mismatch at the border.

[Brief explanation of the drawing]

Figure 1 is a schematic diagram of the system configuration of the present invention, Figure 2 is a diagram showing an example of a component image, Figures 3 (a) and (b) are diagrams showing the original image and area information, and Figure 4 is the image data. Fig. 5 is a diagram showing an example of image synthesis, Fig. 6(a)
, (b) is a diagram explaining area information of a component image, FIG. 7 is a diagram explaining encoded data of a component image, and FIG. 8(a)
, (b) is a diagram explaining encoded data of an image, No. 9
Figures (a) and (b) are diagrams for explaining edge processing, Figure 1θ is a diagram for explaining restored image data, and Figures 11 (a), (b), and (c) are diagrams for explaining conventional image synthesis. Figures 12(a) and 12(b) are diagrams explaining the conventional compression method for component images. Figures 13 and 14 are diagrams explaining the problems of conventional image synthesis. The figure is a diagram illustrating misalignment at boundaries when images are combined. ! ...Color entire background image data, 2.. Part image data, 3.. Image input device, 4..
- Extraction processing section, 5. Irreversible compression section, 6.. Reversible compression section, 7.. Recording and reproducing device, 8.. Irreversible decoding section,
9... Reversible decoding unit, 10... Combination processing unit, 11.
...Image display device. ((++Original drawing 3rd figure (bl inclined carpet edge 4th figure Patent applicant Sharp Co., Ltd. agent)
Patent attorney, above, and 1 other person (G) Figure 6 (b) Figure 8 Chi 3 Items Literary talent in the image - Figures 1 and 9 (a) @No (Yu 1ri) Sodokoro Tamari M
) (b) Synthesis of null 1ri (2...ruri) Fig. 14 Fig. 13 Image weighting Fig. 15

Claims (3)

[Claims] (1) In an image encoding and synthesis method that encodes and saves a component image and the entire background image and synthesizes them, a target image is cut out from the component image and whether each pixel in the component image is included in the target image is determined. Create area information to determine whether the image is included in the background image, and irreversibly compress the color information of each pixel of the entire background image and the color information of each pixel of the target image in the component image. At the same time, the region information of each pixel of the component image is reversibly compressed and stored, and the irreversibly compressed color information is decoded, and the reversibly compressed region information is decoded to create the region information in the component image. Image encoding characterized in that the target image in the component image and the entire background image are combined based on the color information of each pixel of the target image, the area information, and the color information of each pixel of the entire background image. Synthesis method. (2) In the image encoding and synthesis method according to claim 1, the area information includes edge information indicating a boundary between the target image and the background image in the component image, and based on the edge information, , an image encoding and synthesis method characterized in that color information of the boundary portion is created. (3) In the image encoding/synthesizing method according to claim 2, the color information of the pixels in the boundary is combined with the color information of the pixels of the target image near the boundary and the pixels of the entire background image near the boundary. An image coding and synthesis method characterized by creating an image by interpolation based on edge information from color information.