JP2000102014A - Image encoding apparatus and method, and providing medium - Google Patents

Abstract

(57) [Summary] [Problem] To find an accurate motion vector. An evaluation value calculation unit calculates an evaluation value (error) indicating a degree of similarity between a reference block input from a block extraction unit and an inspection block sequentially input from a block extraction unit. , A motion vector corresponding to the difference between the position of the reference block and the position of the test block is calculated. The minimum value register 17 stores the minimum value S of the evaluation value.
_The next register 18 stores the second smallest evaluation value S ₁ , storing the _min and the corresponding motion vector. The difference calculator 19 calculates the difference S _min −S ₁ between the evaluation values. When the determination unit 20 determines that the difference S _min −S ₁ is not larger than the predetermined threshold, the determination unit 20 outputs the information to the control unit 21. The control unit 21 causes the block cutout units 13 and 14 to enlarge the size of the block to be cutout in response to the information from the determination unit 20.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus and method, and a providing medium, and more particularly, to an image coding method capable of obtaining an accurate motion vector by making the size of a reference block variable. And a providing medium.

[0002]

2. Description of the Related Art In a high-efficiency video coding technique represented by the MPEG (Moving Picture Experts Group) scheme,
Motion compensation and discrete cosine transform (DCT: Discrete Cosine Tra)
nsform) is used.

In motion compensation, for example, the current frame shown in FIG. 7A and a reference frame shown in FIG. 7B are compared with each other, and a reference block of the current frame and a reference frame of the current frame are compared. The difference between the position of the reference frame and the position of the test block that best matches the block is detected as a motion vector. Note that such a motion vector detection method is called block matching.

[0004]

In the above-described block matching, if the reference block of the current frame is set small, a large number of test blocks matching the reference block may be present in the reference frame. There was a problem that could not be obtained accurately.

The present invention has been made in view of such circumstances, and it is an object of the present invention to make it possible to obtain an accurate motion vector by making the size of a reference block variable.

[0006]

According to a first aspect of the present invention, there is provided an image encoding apparatus comprising: a first extracting unit for extracting a reference block from a current frame; and a second extracting unit for extracting a check block from a reference frame. Error calculating means for calculating an error between the reference block extracted by the first extracting means and the test block extracted by the second extracting means, and a reference block extracted by the first extracting means and the second extracting means. Detecting means for detecting a motion vector from the position of the extracted test block; difference calculating means for calculating a difference between a minimum value of the error calculated by the calculating means and the next smallest value; and calculation results of the difference calculating means. And changing means for changing the sizes of the reference block and the inspection block.

According to a second aspect of the present invention, in the image coding method, a first extraction step of extracting a reference block from a current frame, a second extraction step of extracting a check block from a reference frame, and a first extraction step An error calculation step for calculating an error between the reference block extracted in step 2 and the inspection block extracted in the second extraction step, and a position of the reference block extracted in the first extraction step and the position of the inspection block extracted in the second extraction step A detection step of detecting a motion vector from a motion vector, a difference calculation step of calculating a difference between a minimum value of an error calculated in the calculation step and a next smaller value, and a reference block and an inspection based on a calculation result of the difference calculation step. Changing the size of the block.

According to a third aspect of the present invention, in the providing medium, a first extraction step of extracting a reference block from a current frame, a second extraction step of extracting a check block from a reference frame, and a first extraction step Calculating an error between the reference block extracted and the test block extracted in the second extraction step, and moving from the position of the reference block extracted in the first extraction step and the position of the test block extracted in the second extraction step A detection step of detecting a vector; a difference calculation step of calculating a difference between a minimum value of an error calculated in the calculation step and a next smaller value; and a calculation result of the reference block and the inspection block based on a calculation result of the difference calculation step. A computer-readable program that causes the image encoding device to execute a process including a changing step of changing the size. And providing a gram.

In the image encoding apparatus according to the first aspect, the image encoding method according to the second aspect, and the providing medium according to the third aspect, a reference block is extracted from a current frame, and an inspection is performed from a reference frame. Blocks are extracted. An error between the reference block and the test block is calculated, and a motion vector is detected from the positions of the reference block and the test block. The difference between the minimum value of the error and the next smaller value is calculated, and the sizes of the reference block and the inspection block are changed based on the calculation result.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below. In order to clarify the correspondence between each means of the invention described in the claims and the following embodiments, each means is described. When the features of the present invention are described by adding the corresponding embodiment (however, an example) in parentheses after the parentheses, the result is as follows. However, of course, this description does not mean that each means is limited to those described.

The image coding apparatus according to the first aspect of the present invention includes a first extracting means (for example, a block extracting unit 13 in FIG. 2) for extracting a reference block from a current frame and a first extracting means for extracting a check block from a reference frame. 2 extracting means (for example, the block extracting unit 14 in FIG. 2) and an error calculating means (for example, an error calculating means for calculating an error between the reference block extracted by the first extracting means and the test block extracted by the second extracting means) Evaluation value calculation unit 15 in FIG. 2 and detection means for detecting a motion vector from the positions of the reference block extracted by the first extraction means and the test block extracted by the second extraction means (for example, the evaluation value of FIG. 2) A calculating unit 15), a difference calculating unit that calculates a difference between the minimum value of the error calculated by the calculating unit and the next smaller value (for example, the difference calculating unit 19 in FIG. 2), and a calculation result of the difference calculating unit. Based on, changing means for changing the size of the reference block and the inspection block (e.g., the control unit 21 of FIG. 2), characterized in that it comprises a.

An example of the configuration of an encoder to which the present invention is applied will be described with reference to FIG. In the encoder 1, sequentially inputted image data (frame data constituting a moving image) is supplied to the subtracter 1 and the motion vector detection circuit 9. The subtracter 1 calculates a difference between the input image data and the predicted image data supplied from the motion compensation circuit 8, and calculates a result of the calculation by a discrete cosine transform (DC).
T) Output to circuit 2. The discrete cosine transform circuit 2 performs a discrete cosine transform on the input difference, and outputs the obtained DCT coefficients to the quantization circuit 3.

The quantizing circuit 3 quantizes the input DCT coefficient, transmits the quantized DCT coefficient to a transmission path (not shown), and outputs the quantized DCT coefficient to the inverse quantization circuit 4. The inverse quantization circuit 4 includes the quantization circuit 3
, And performs inverse quantization corresponding to the quantization of the quantization circuit 3, and outputs the result to the inverse DCT circuit 5. The inverse DCT circuit 5 converts the DCT input from the inverse quantization circuit 4
An inverse DCT process corresponding to the discrete cosine transform process of the discrete cosine transform circuit 2 is performed on the coefficient, and the obtained difference data is output to the adder 6.

The adder 6 restores the original image data by adding the predicted image data input from the motion compensation circuit 8 and the difference data supplied from the inverse DCT circuit 5, and restores the original image data. Output to Frame memory 7
Stores the restored image input from the adder 6.

The image data stored in the frame memory 7 is read out by the motion compensating circuit 8 and subjected to motion compensation based on the motion vector, and then supplied to the subtractor 1 and the adder 6 as predicted image data.

The motion vector detection circuit 9 detects a motion vector of the input image data (frame image),
Output to the motion compensation circuit 8.

FIG. 2 shows a detailed configuration example of the motion vector detection circuit 9 of FIG. The current frame memory 11
When image data (frame data) is input, the stored image data is supplied to the reference frame memory 12 and the currently input image data is stored. When the image data is input from the current frame memory 11, the reference frame memory 12 stores the image data. That is, in the current frame memory 11 and the reference frame memory 12, a current frame or a reference frame which is temporally preceding or succeeding is stored.

The block extracting section 13 converts the current frame stored in the current frame memory 11 into a control section 21.
The reference block is cut out in accordance with the block size and the signal designating the position input from the control unit and output to the evaluation value calculation unit 15. The block cutout unit 14 cuts out test blocks sequentially from the reference frame stored in the reference frame memory 12 in accordance with the block size input from the control unit 21 and the signal designating the position, and outputs the check blocks to the evaluation value calculation unit 13. Output.

The evaluation value calculation unit 15 calculates an evaluation value (error) indicating the degree of similarity between the reference block input from the block extraction unit 13 and the inspection block sequentially input from the block extraction unit 14 according to the following equation (1). ) And outputs the result to the minimum value detection unit 16. Evaluation value = ΣΣ | Y _ij −X _ij | (1) where X _ij is the pixel value of the reference block, and Y _ij is the pixel value of the inspection block.

The evaluation value calculation unit 15 calculates a motion vector corresponding to the difference between the position of the reference block and the position of the inspection block, and outputs the calculated motion vector to the minimum value detection unit 16.

The minimum value detection unit 16 compares the evaluation value input from the evaluation value calculation unit 15 with the evaluation value stored in the minimum value register 17, and determines the smaller comparison value and the corresponding motion. The vector is stored (updated) in the minimum value register 17. When the stored evaluation value is updated by the minimum value detection unit 16, the minimum value register 17 outputs the evaluation value stored so far to the next-point register 18. Therefore, finally, the minimum value register 17 stores the minimum value S _min of the evaluation value and the corresponding motion vector, and the next point register 18 stores the second smallest evaluation value S ₁ . .

The difference calculating section 19 reads the evaluation value S _min and the evaluation value S ₁ from each of the minimum value register 17 and the next point register 18, calculates the difference S _min −S ₁ therebetween, and Output.

The determining unit 20 determines whether the difference S _min -S ₁ of the evaluation value input from the difference calculating unit 19 is larger than a predetermined threshold, and determines whether the difference S _min -S ₁ is larger than the predetermined threshold. If the information is not large, the information is output to the control unit 21. When the determination unit 20 determines that the difference S _min −S ₁ is larger than the predetermined threshold, the determination unit 20 outputs a corresponding motion vector to the subsequent motion compensation circuit 8.

The control section 21 outputs a signal designating the position of the reference block to the block cutout section 13 and sends the signal to the block cutout section 14 in order to search for an inspection block matching the reference block. , The position of the inspection block is sequentially changed within a predetermined search range of the reference frame, and a signal for designating the position is output. Further, the control unit 21 outputs a signal for enlarging the size of the block to be cut out to the block cutout units 13 and 14 in accordance with the information from the determination unit 20.

Next, the motion vector detection processing of the motion vector detection circuit 9 will be described with reference to the flowchart of FIG. This process is started when the current frame or the reference frame that is temporally preceding or succeeding is stored in the current frame memory 11 and the reference frame memory 12.

In step S1, the control unit 21 controls the block cutout units 13 and 14 as shown in FIG.
, The parameter n of the signal specifying the size ((2n + 1) × (2n + 1) pixels) of the reference block and the inspection block to be output is initialized to 2. Further, the control unit 21 outputs a signal specifying the position of the reference block to the block cutout unit 13, and outputs a signal specifying the position of the inspection block by sequentially changing the position to the block cutout unit 14.

In step S2, for example, as shown in FIG. 5A, the block cutout unit 13 converts the current frame supplied from the current frame memory 11 into the control unit 21.
A reference block is cut out in accordance with the signal input from, and output to the evaluation value calculation unit 13. For example, as shown in FIG. 5B, the block cutout unit 14 converts the control frame from the reference frame supplied from the reference frame memory 12 into the control unit 2.
Inspection blocks are sequentially cut out according to the signal input from 1 and output to the evaluation value calculation unit 13.

The evaluation value calculation unit 15 calculates an evaluation value indicating the degree of similarity between the reference block input from the block extraction unit 13 and the inspection block sequentially input from the block extraction unit 14 using Expression (1). At the same time, the motion vector corresponding to the difference between the position of the reference block and the position of the inspection block is calculated and output to the minimum value detection unit 16.

In step S3, the minimum value detector 16
Compares the evaluation value input from the evaluation value calculation unit 15 with the evaluation value read from the minimum value register 17, and stores (updates) the smaller comparison value and the corresponding motion vector in the minimum value register 17. ). Minimum value register 1
7, the evaluation value stored so far is the minimum value detection unit 16;
When it is updated by (1), the evaluation value stored so far is output to the next register 18. Therefore, finally, the minimum value register 17 stores the minimum value S _min of the evaluation value and the corresponding motion vector.
The 8, smaller evaluation values S ₁ to the second is stored.

In step S4, the difference calculation section 19
Reads the evaluation value S _min or the evaluation value S ₁ from each of the minimum value register 17 and the next-point register 18,
The difference S _min −S ₁ is calculated and output to the determination unit 20. The determination unit 20 determines that the difference S _min −S ₁ between the evaluation values input from the difference calculation unit 19 is a predetermined threshold (for example, if the pixel value is 8
If the value is represented by bits and the maximum value is 255, it is determined whether or not the value is larger than twice (2 (2n + 1) ² ) the number of pixels in the block. When it is determined that the difference S _min −S ₁ is not larger than the predetermined threshold, that is, when there is no difference between a plurality of test blocks that are candidates that match the reference block in FIG. .

In step S5, the judgment unit 20 outputs the judgment result in step S4 to the control unit 21, and the control unit 2
1 corresponds to the determination result, and the block cutout unit 1
The parameter n of the signal designating the size ((2n + 1) × (2n + 1) pixels) of the reference block and the test block to be output to 3, 14 is incremented by one. Correspondingly, the size of the block cut out by the block cutout units 13 and 14 is enlarged as shown in FIG.

Then, in step S4, the difference S_min-S₁
Until it is determined that is greater than a predetermined threshold.
The processing of S2 to S5 is repeated, and in step S4, the difference
Minute S _min-S₁Is determined to be greater than a predetermined threshold.
In other words, the inspection that matches the reference block in FIG.
If there is only one block, the process proceeds to step S6.

In step S 6, the judgment section 20 outputs a motion vector corresponding to the evaluation value S _min to the subsequent motion compensation circuit 8.

As described above, in the present embodiment, when there are a plurality of test blocks similar to the reference block, the size of the reference block is enlarged to narrow the plurality of test blocks to one. Therefore, an accurate motion vector can be detected without unnecessarily increasing the number of operations.

It should be noted that the present invention is applied to an image conversion apparatus for changing a position shape of a class tap and a prediction tap in accordance with a motion vector, classifying the class, and converting a low-resolution image into a high-resolution image. Is also possible.

The computer program for performing each of the above processes can be provided to the user via a network medium such as the Internet or a digital satellite in addition to a medium such as a magnetic disk or a CD-ROM. it can.

[0037]

As described above, according to the image encoding apparatus of the first aspect, the image encoding method of the second aspect, and the providing medium of the third aspect, the reference block and the check block are used. The size of the reference block and the size of the test block are changed based on the difference between the minimum value of the error and the next smaller value, so that an accurate motion vector can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an encoder 1 to which the present invention has been applied.

FIG. 2 is a block diagram illustrating a detailed configuration example of a motion vector detection circuit 9 in FIG. 1;

FIG. 3 is a flowchart illustrating a motion vector detection process of a motion vector detection circuit 9;

FIG. 4 is a diagram illustrating the sizes of a reference block and an inspection block.

FIG. 5 is a diagram illustrating a motion vector detection process.

FIG. 6 is a diagram illustrating a motion vector detection process.

FIG. 7 is a diagram illustrating a motion vector.

[Explanation of symbols]

9 motion vector detection circuit, 11 current frame memory, 12 reference frame memory, 13, 14 block cutout section, 15 evaluation value calculation section, 16 minimum value detection section, 17 minimum value register, 18 order register, 19 difference calculation section, 20 judgment part, 21 control part

Claims (3)

[Claims] 1. An image encoding apparatus for detecting a motion vector from a current frame and a reference frame which are temporally preceding and succeeding, comprising: a first extraction unit for extracting a reference block from the current frame; A second extraction unit to be extracted; a reference block extracted by the first extraction unit;
Error calculating means for calculating an error with respect to the inspection block extracted by the extracting means, and a reference block extracted by the first extracting means;
Detecting means for detecting a motion vector from the position of the test block extracted by the extracting means, a difference calculating means for calculating a difference between a minimum value of the error calculated by the calculating means and a next smaller value, and the difference calculation An image coding apparatus comprising: a changing unit configured to change a size of the reference block and the size of the inspection block based on a calculation result of the unit. 2. An image coding method for detecting a motion vector from a current frame and a reference frame that are temporally preceding and succeeding, comprising: a first extraction step of extracting a reference block from the current frame; A second extraction step for extracting; an error calculation step for calculating an error between the reference block extracted in the first extraction step and the test block extracted in the second extraction step; A detecting step of detecting a motion vector from the extracted reference block and the position of the test block extracted in the second extracting step; calculating a difference between a minimum value of the error calculated in the calculating step and a next smaller value A difference calculation step, and based on the calculation result of the difference calculation step, the reference block and the inspection block. Changing the size of the lock. 3. An image encoding apparatus for detecting a motion vector from a current frame and a reference frame that are temporally preceding and succeeding, a first extraction step of extracting a reference block from the current frame, and a check block from the reference frame. A second extraction step for extracting; an error calculation step for calculating an error between the reference block extracted in the first extraction step and the test block extracted in the second extraction step; A detecting step of detecting a motion vector from the extracted reference block and the position of the test block extracted in the second extracting step; calculating a difference between a minimum value of the error calculated in the calculating step and a next smaller value A difference calculation step, based on a calculation result of the difference calculation step, the reference block and the inspection block Providing a computer-readable program for executing a process including a changing step of changing the size of a computer.