JP2000217108A - Image-coding method and storage medium thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize substantial high image quality transmission by transmitting the image of a target part with higher image quality than those of others, even in the case of transmission medium with a low transmission rate. SOLUTION: A discrete cosine transform(DCT) circuit 10 applies discrete cosine transformation to a received image. A quantization device 12 quantizes a DCT coefficient outputted from the DCT circuit 10, and a variable length coding circuit 14 applies variable length coding (e.g. Huffman coding) to the output of the quantization device 12. The output of the variable length coding circuit 14 is coded image information and is fed to a transmission line or a storage medium, for example, a quantization control circuit 16 controls the quantization characteristic of the quantization device 12 differently for a target part and other parts, according to the information of a representative quantization step, the center coordinate of a macro block, the coordinate of the target point and other information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to an image encoding method and a storage medium.

[0002]

2. Description of the Related Art In a conventional image coding method, the inside of a screen is coded under uniform coding conditions, and coding is not performed while distinguishing a range of interest on an image from other ranges. .

[0003] In the coding of the image pickup signal, no consideration is given to a change in the image pickup range of the camera.

[0004]

In the moving image transmission, if a specific portion in a screen, for example, a portion that the image receiver wants to pay attention to can be transmitted with high image quality and other portions can be transmitted with low image quality, it is limited. Although the transmission rate obtained can be effectively used, such a technique cannot be realized by the conventional technology. .

In the moving picture coding method using the difference between frames, a high quality image is gradually obtained under certain conditions such as when the image does not change greatly. There is a demand for higher image quality more quickly than other parts, but the conventional technology cannot cope with it.

[0006] Even in the case of transmitting a still image having a larger transmission data amount than the transmission capacity, there is a demand that a specific portion has higher image quality than other portions, but this cannot be realized by the conventional technology.

In a live camera system in which a video receiver can control the camera attitude via a network, the subject to be viewed has a higher image quality than other parts even though the camera attitude is controlled so as to face the object to be viewed. And there is a problem that the image quality does not quickly become high.

In a system for transmitting an image of a camera having an object tracking function, there is a demand for viewing a tracked object and its surroundings with high image quality. However, in the conventional example,
The tracked object and its surroundings have the same image quality as the other parts, and the image quality does not quickly become high prior to the other parts.

In a video transmission system for transmitting an image of a camera whose imaging range can be freely changed via a transmission path having a relatively small transmission capacity, it is desired to view an image whose imaging range has been changed at an early stage. There is a problem that an image in which the imaging range is being changed is transmitted, and as a result, an image in which the imaging range has been changed cannot be immediately viewed.

An object of the present invention is to provide an image coding method which solves these problems.

It is another object of the present invention to provide an image encoding method that enables a specific portion on an image to be transmitted with higher image quality than other portions.

It is another object of the present invention to provide an image coding method for transmitting a specific portion on an image so that the quality of the image is quickly increased as compared with the other portions in moving image transmission.

According to the present invention, when transmitting a still image having a large transmission data amount compared to the transmission path capacity, an image to be transmitted so that a specific portion on the image becomes higher in image quality more quickly than other portions. The purpose is to present an encoding method.

According to the present invention, in a live camera system in which a video receiver can control a camera attitude via a network, a subject to be viewed is transmitted so as to have high image quality or to quickly obtain high image quality as compared with other portions. It aims to present an image coding method.

The present invention is also directed to a system for transmitting an image of a camera having an object tracking function so that a tracked object and its surroundings can have a high image quality or a high image quality as compared with other parts. The purpose of the present invention is to present an image encoding method to be transmitted to a computer.

According to the present invention, in an image transmission system for transmitting an image of a camera whose imaging range can be freely changed by using a transmission line having a relatively small capacity, an image whose imaging range has been changed can be quickly viewed. It aims to present an image coding method.

Another object of the present invention is to provide a storage medium for storing program software for executing the above-described image encoding method.

[0018]

An image encoding method according to the present invention is characterized in that a larger amount of encoded data is preferentially allocated to a target portion of an image than other portions.

The image encoding method according to the present invention is further characterized in that a larger amount of encoded data is preferentially allocated than other parts until the image quality of a target part of the image reaches a predetermined level. .

The image encoding method according to the present invention is characterized in that the amount of encoded data is reduced while the imaging range is changing.

The storage medium according to the present invention stores the program and software of the above-described image encoding method.

[0022]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a schematic block diagram showing an embodiment of the present invention. Discrete cosine transform (DCT) circuit 10
Performs a discrete cosine transform of an input image. The quantizer 12 quantizes the DCT coefficient output from the DCT circuit 10,
The variable length coding circuit 14 performs variable length coding (for example, Huffman coding) on the output of the quantization circuit 12. The output of the variable length encoding circuit 14 is encoded image information, and is supplied to, for example, a transmission path or a storage medium.

The quantization control circuit 16 quantizes the quantization characteristic of the quantizer 12 between the target portion and the other portions according to the representative quantization step information, the center coordinates of the macroblock, the coordinates of the target point, and other information. Is controlled as follows.

The image coding apparatus shown in FIG. 1 is, for example, in the case of moving picture coding, for example, according to ITU-T Recommendation H.264. Assume that the image is encoded according to H.263.

FIG. 2 shows a sub-QCIF image. Macroblock boundaries are shown for clarity. The mark x in FIG. 2 is taken as the point of interest. The noticed portion is a peripheral portion including the noticed point.

In this embodiment, an image is coded so that more coded data is allocated to a macro block of a target portion. Specifically, for example, the quantization control circuit 16 makes the quantization step for the macroblock including the point of interest slightly smaller than the typical value. Of course, then
In order to keep the code amount of the entire screen constant, the quantization control circuit 16 may make the quantization step for other macroblocks slightly larger than the typical value.

When such a function of the quantization control circuit 16 is determined by fuzzy operation, as a rule # 1, the quantization step is set slightly smaller than the candidate value in the portion of interest.
As rule # 2, the quantization step is made slightly larger than the candidate value for the part other than the part of interest. FIG.
Shows an example of the characteristics of the membership function representing the attention portion.

A method of calculating a quantization step applied to each macroblock will be briefly described. For example, in the macroblock 20 in FIG. 2, the Euclidean distance between the center coordinates and the point of interest is (76 ² +12 ² ) ^1/2 = 76.9. When this is applied to the graph of FIG. 3, the degree of belonging of the macro block 20 to the target portion becomes 0. When this value is applied to the previous rule # 1, rule # 1 does not hold because the degree of belonging to the target portion is 0.

In rule # 2, if 1-0 = 1 is used as the negation of the noticed part, rule # 2 is established and macro block 20
, The result is that the quantization step is made slightly larger than the candidate value.

If the sum operation is used as a result of combining rule # 1 and rule # 2, the quantization step for macroblock 20 is slightly larger than the candidate value as the final result.

For example, assume that 8 is given as a quantization step candidate value input to the quantization control circuit 16. If the operation of “increase slightly” is interpreted as “increase by 10%”, then 8 × 1.1 = 8.8, and if the integer processing is rounded off, the quantization step is set to 9. Become.

When the Euclidean distance between the macroblock 22 and the point of interest is calculated, (12 ² +4 ² ) ^1/2 = 12.6. When this value is applied to FIG. 3, the degree of belonging to the target portion is 0.68. This is called Rule # 1 and Rule # 2
Will be applied.

The operation of "decreasing slightly" is changed to "20%
When interpreted as “reduce”, in rule # 1, 0.68 × 0.2 = 0.14, and in rule # 2, (1−0.68) × 0.1 = 0.03. This means that the quantization step is reduced by 14% and increased by 3%. As a result, 8 × (1 + (− 0.14 + 0.03) = 7.12 ”is obtained. The step becomes 7.

It is not necessary that the number of attention portions is one. If there are a plurality of attention portions, the degrees of belonging to the attention portion are calculated for all the attention points, and the logical sum of them is added to the attention portion. The degree of affiliation may be used.

The fuzzy rule when the image quality is also taken into consideration may be as follows. As rule # 1, the quantization step is made slightly smaller than the candidate value for the portion of interest and not the desired image quality. Rule # 2
The quantization step is made slightly larger than the candidate value in a portion other than the portion of interest or in a portion having a desired image quality. FIG. 4 shows an example of a membership function representing a desired image quality.
Shown in

In the moving picture coding method, an inter-frame difference method is used to increase the compression ratio. It is assumed that frame images α and β are continuous and are encoded in this order. Assuming that a decoded image of the image α is an image α ′, in the inter-frame difference encoding, a difference image between the image α ′ and the image β is encoded. The macro blocks 20 and 22 in the screen example shown in FIG. 2 will be described as an example. The correlation between the image α ′ and the image β in the macro block 20 is obtained. Assuming that this value is 0.5, as shown in FIG.
Is 0.77. Assuming that the intersection set in rule # 1 is the minimum value operation and the union set in rule # 2 is the maximum value operation, min (0, 1-0.77) = 0 max (1-0, 0.77) = 1 Therefore, according to rule # 2, the quantization step is set to 1 × 1.1 = 1.1, and if the quantization step candidate value is set to 8, the quantization step to be obtained is 8 × 1.1 = 8.8
When rounded to an integer, the result is 9.

Needless to say, the above embodiment can be used not only for encoding a moving image but also for encoding a still image.

It is assumed that the portion of interest is the central portion of the image. This is because the central part of the image is generally the most noticeable part, and the effect of increasing the image quality is high.

In photographing a moving object, a method of automatically tracking a subject is known. In a camera employing such an automatic tracking method, a subject to be tracked is set as a target portion, so that the subject can be transmitted with higher image quality than others.

Other information to be input to the quantization control circuit 16 includes, for example, an imaging range parameter (parameters such as pan, tilt, and zoom) of the camera. FIG.
Indicates a membership function indicating a change in the imaging range according to the panning speed. The vertical axis represents the degree of belonging, and the horizontal axis represents the panning speed. In this case, as a fuzzy rule to be added, the quantization step is increased with respect to a change in the imaging range. Furthermore, a rule may be added to prevent encoded data from being generated at all when the movement becomes intense to some extent.

[0042]

As can be easily understood from the above description, according to the present invention, a specific portion of an image can be designated and encoded so as to have higher image quality than other portions or to achieve higher image quality more quickly. . Therefore, even with a transmission medium having a low transmission rate, the target portion can be transmitted with higher image quality than the others, and substantially high image quality transmission can be realized.

When the present invention is applied to a live camera system in which a video receiver can control a camera attitude via a network, a subject to be viewed can have higher image quality than other portions.

If the present invention is applied to a system for transmitting an image of a camera having an object tracking function, a tracked object (and its surroundings) can be transmitted with higher image quality than other parts.

When the present invention is applied to a system for transmitting an image of a camera whose imaging range can be freely changed by using a low-capacity transmission line, an image whose imaging range has been changed can be quickly viewed.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of an embodiment of the present invention.

FIG. 2 is an example of a screen configuration of a sub-QCIF used to explain the operation of the present embodiment.

FIG. 3 is a characteristic diagram of a membership function representing a portion of interest.

FIG. 4 is a characteristic diagram of a membership function representing a desired image quality.

FIG. 5 is a characteristic diagram of a membership function representing an imaging range change.

[Explanation of symbols]

 10: discrete cosine transform (DCT) circuit 12: quantizer 14: variable length coding circuit 16: quantization control circuit

Claims (8)

[Claims] 1. An image encoding method characterized in that a larger amount of encoded data is preferentially allocated to a target portion of an image than other portions. 2. An image encoding method, wherein a larger amount of encoded data is preferentially allocated than other parts until the image quality of a target part of the image reaches a predetermined level. 3. The image encoding method according to claim 2, wherein the image is a still image. 4. The image encoding method according to claim 1, wherein the target portion is a central portion of the image. 5. The moving picture coding method according to claim 1, wherein the target portion is centered on an object that is automatically tracked. 6. An image encoding method, wherein the amount of encoded data is reduced while the imaging range is changing. 7. The image encoding method according to claim 6, wherein encoded data is not generated while the imaging range is changing. 8. A storage medium storing the program software of the image encoding method according to claim 1, 2 or 6.