JP4160513B2 - Moving image luminance change parameter estimation method, moving image luminance change parameter estimation program and recording medium thereof, and moving image encoding apparatus, moving image encoding method, moving image encoding program and recording medium thereof - Google Patents

Description

  The present invention relates to a technique for estimating a luminance change amount between image frames required in an encoding method for efficiently transmitting and storing digital moving images, an image processing method for moving image editing, and the like. The present invention relates to a technique for encoding a moving image using a quantity estimation technique.

As a conventional technique for estimating a global luminance change amount between moving image frames, for example, there is a technique described in Patent Document 1. In Patent Document 1,
x ′ = W · x + C (1)
The change in luminance is expressed by the equation. In Equation (1), x is the luminance of the pixel before the luminance change, and x ′ is the luminance of the pixel after the luminance change. W and C are constants and are parameters representing the amount of change in luminance. This is called a luminance change parameter.

  As can be seen from equation (1), estimating the amount of change in luminance between one image frame (hereinafter referred to as image frame α) and another image frame (hereinafter referred to as image frame β) means that image frame α Is equivalent to calculating the luminance change parameter between the image frame β and the image frame β.

In the technique described in Non-Patent Document 1, two types of brightness change parameters are prepared between the image frame α and the image frame β in order to cope with the case where the brightness change varies depending on the location in the image frame. , You can select either for each block.
Japanese Patent No. 2938412 "Adaptive Weighted Motion Compensation in MPEG-4 AVC", 2003 Image Coding Symposium (PCSJ2003), P-5.14, pp.89-90

  In the technique of Non-Patent Document 1, as described above, two types of brightness change parameters between the image frame α and the image frame β are prepared, and either one can be selected for each block. However, since there is no appropriate calculation method for the two types of luminance change parameters, W is set to W = 1 for both types of parameters. For C, the average value of the entire screen of the encoding target image frame and the two types of reference image frames is merely used.

  Therefore, this method does not necessarily use parameters that represent the actual amount of change in luminance. For example, it cannot be said that sufficient encoding efficiency is obtained even if this method is applied to image coding. There was a problem.

  In the present invention, in order to solve the above problems, even when a plurality of luminance changes occur, it is possible to estimate the luminance change with high accuracy by appropriately calculating the luminance change parameter. Accordingly, an object of the present invention is to provide a technique capable of performing encoding more efficiently than in the past.

  In the present invention, in order to solve the above-described problem, first, the image frame β is divided into N (N ≧ 2) regions including one or more blocks, and for each region, a corresponding region of the image frame α. The brightness change parameter is calculated between Next, an optimal one is selected from the N types of luminance change parameters for each block of the image frame β. After selecting appropriate parameters for all the blocks, a set of blocks with the same luminance change parameter selected is set as a new area, and the luminance change parameter is recalculated for each of the new N areas.

  The above process is repeated, and when the value of the luminance change parameter is almost unchanged, the finally obtained N types of luminance change parameters are set as the luminance change parameters of the image frame β.

  In the present invention, as a method of estimating a luminance change amount between two image frames, a method of calculating a luminance change parameter suitable for each region while classifying the region according to the local property of the image frame is used. Therefore, it is possible to accurately estimate the luminance change even when a plurality of luminance changes occur.

  In addition, by using such a luminance change estimation method in an image encoding device, it is possible to perform encoding more efficiently than in the past.

  FIG. 1 is a flowchart of the moving image luminance change parameter estimation process according to the embodiment of the present invention. FIG. 2 is a diagram showing a configuration of an image frame in the present embodiment. FIG. 3 is a diagram showing an example of the region (initial state) of the image frame β in the present embodiment. FIG. 4 is a diagram showing an example of the region (reset state) of the image frame β in the present embodiment. Hereinafter, the flow of luminance change parameter estimation processing of a moving image in the present embodiment will be described with reference to FIGS.

  In the present embodiment, a case where the number N of luminance change parameters is 4 will be described as an example. Further, as shown in FIG. 2, the size of the image frame is 320 horizontal pixels × 256 vertical pixels, and the size of each block of the image frame is 16 horizontal pixels × 16 vertical pixels. Therefore, the image frame is composed of 20 horizontal blocks × 16 vertical blocks. Of course, these are examples, and the present invention is not limited to these implementations.

  Step S1: A sufficiently large value is substituted for the variable D.

  Step S2: The image frame β is divided into four, and regions (regions A to D) are set. Here, as shown in FIG. 3, it is assumed that all regions are divided so as to be equal. One area is composed of 10 horizontal blocks × 8 vertical blocks.

  Step S3: For each set region, a luminance change parameter between the corresponding portion (region) of the image frame α is calculated. Here, the calculation of the luminance change parameter of each area is realized by, for example, applying the technique for calculating the luminance change parameter of the entire screen described in Patent Document 1 to each area instead of the entire screen. can do. As a result, four types of luminance change parameters (luminance change parameters A to D) are obtained.

  Step S4: For each block of the image frame β, one of the four types of luminance change parameters calculated in Step S3 is selected. As a procedure for selecting the luminance change parameter for each block, first, for each of the four types of luminance change parameters, the luminance change using the equation (1) is calculated from the luminance value x of the corresponding portion (block) of the image frame α. A later luminance value x ′ is obtained. Next, an error between the luminance value x ′ after the luminance change obtained by the equation (1) and the actual luminance value of the image frame β is obtained, and the absolute value sum in the block of the obtained error is obtained as four types of luminance change parameters. Calculate for each. The luminance change parameter that minimizes the sum of the absolute values of the calculated errors in the block is selected as the luminance change parameter of the block.

  Step S5: The sum total of the entire image frame β of the sum of absolute values of the errors of the respective blocks calculated in step S4 is obtained, and it is defined as d. Here, as the in-block absolute value sum of the error of each block used for calculating the total sum of the entire image frame β, the in-block absolute value sum of the error calculated using the selected luminance change parameter is used.

  Step S6: If D-d> e, proceed to Step S7. If D−d ≦ e, the process proceeds to step S9. Here, e is a positive value set in advance. The determination at step S6 is the d newly obtained by the processing at step S3 to step S5 in the area reset at the next step S7 and the previous d (substituted for D at step S8). When the difference between the two is less than or equal to the threshold value e, that is, even if the region is reset and the processing in steps S3 to S5 is executed, the total sum d of the entire image frame β of the sum of absolute values of the errors in each block is not much This is a determination for terminating the process when the change has stopped.

  Step S7: Each block of the image frame β is classified for each selected luminance change parameter, and set as a new area. For example, as shown in FIG. 4, a block whose luminance change parameter is the luminance change parameter A is reset as a region A ′. Similarly, the blocks whose luminance change parameters are the luminance change parameters B to D are also set as areas B 'to D', respectively.

  Step S8: Substitute the value of d into D, and return to step S3.

  Step S9: The four areas finally set in step S7 and the four types of luminance change parameters finally calculated in step S3 are determined as the areas and luminance change parameters in the image frame β, and processed. Exit.

  FIG. 5 is a diagram illustrating a configuration example of the moving picture coding apparatus according to the embodiment of the present invention. The moving image encoding apparatus shown in FIG. 5 performs inter-frame prediction of a moving image using the above-described luminance change parameter estimation processing, and encodes a prediction error signal.

  An encoding target image frame is input from the input terminal 1. On the other hand, the frame memory 7 stores reference image frames encoded one frame before.

  The luminance change parameter estimation unit 9 executes the above-described luminance change parameter estimation processing. That is, the encoding target image frame input from the input terminal 1 corresponds to the image frame β, and the reference image frame stored in the frame memory 7 corresponds to the image frame α. For example, the luminance change parameter estimation unit 9 calculates four types of luminance change parameters between the two images by the luminance change parameter estimation process shown in FIG. Information indicating which of the four types of luminance change parameters to use for which block of the image frame to be encoded is sent to the luminance change compensator 8 and the code assigner 10.

  The luminance change compensation unit 8 compensates the luminance change for the reference image frame stored in the frame memory 7 by the luminance change parameter sent from the luminance change parameter estimation unit 9. The code assigning unit 10 assigns a code to the information sent from the luminance change parameter estimating unit 9 and sends it out from the output terminal 12.

  The signal of the reference image frame compensated by the luminance change compensation unit 8 is a prediction signal for the signal of the encoding target image frame, and the prediction error signal obtained by the subtracter 2 is quantized by the quantization unit 3. The quantized signal is sent to the code allocation unit 4 and the inverse quantization unit 5.

  The code assigning unit 4 assigns a code to the quantized signal sent from the quantizing unit 3 and sends it out from the output terminal 11. The inverse quantization unit 5 inversely quantizes the quantized signal sent from the quantization unit 3 to obtain a prediction error signal after quantization. The quantized prediction error signal is added to the signal of the reference image frame compensated by the luminance change compensator 8 by the adder 6 and stored in the frame memory 7 as a predicted reference image for the next encoding target frame.

  Here, in the moving image coding according to the present embodiment, in order to explain the gist of the present invention in an easy-to-understand manner, an example in which only luminance change compensation is performed has been described. Needless to say, the moving picture coding according to the present invention can be implemented in combination with motion compensation or other coding methods.

  The process for estimating the luminance change parameter of the moving image and the process of encoding the moving image described above can be realized by a computer and a software program. The program can be recorded on a computer-readable recording medium, It is also possible to provide through.

It is a brightness | luminance change parameter estimation process flowchart of the moving image in embodiment of this invention. It is a figure which shows the structure of the image frame in this Embodiment. It is a figure which shows the example of the area | region (initial state) of the image frame (beta) in this Embodiment. It is a figure which shows the example of the area | region (state reset) of the image frame (beta) in this Embodiment. It is a figure which shows the structural example of the moving image encoder in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Input terminal 2 Subtractor 3 Quantization part 4,10 Code allocation part 5 Inverse quantization part 6 Adder 7 Frame memory 8 Luminance change compensation part 9 Luminance change parameter estimation part 11, 12 Output terminal

Claims (8)

A method for estimating a luminance change parameter representing a luminance change amount between image frames in a moving image,
A first process of dividing an image frame into N (N ≧ 2) regions of one or more blocks;
A second process of obtaining N types of brightness change parameters by calculating brightness change parameters for each of the divided areas;
A third step of selecting an optimal one of N types of luminance change parameters for each block of the image frame;
A fourth step of classifying blocks of the image frame for each selected luminance change parameter, and setting a set of blocks with the same luminance change parameter as new N regions;
And a fifth process of acquiring new N types of brightness change parameters by recalculating the brightness change parameters for each new region. The method for estimating a luminance change parameter of a moving image according to claim 1,
Repeat the third process to the fifth process until a predetermined condition is satisfied,
A method of estimating a luminance change parameter of a moving image, wherein N types of luminance change parameters when the condition is satisfied are determined as finally estimated luminance change parameters. Claim 1 or the luminance variation of the moving picture for performing a luminance variation parameter estimation method of a moving image according to a computer to claim 2 parameter estimation program. Moving image brightness variation parameter estimation program recorded a computer-readable record medium for executing the luminance variation parameter estimation method on the computer of a moving image according to claim 1 or claim 2. A video encoding device that performs inter-frame prediction of a video by luminance change compensation using a luminance change parameter between image frames and encodes a prediction error signal,
A moving picture encoding apparatus comprising: means for estimating a luminance change parameter for the luminance change compensation using the moving picture luminance change parameter estimation method according to claim 1. A video encoding method that performs inter-frame prediction of a video by luminance change compensation using a luminance change parameter between image frames and encodes a prediction error signal,
A method for encoding a moving image, comprising: estimating a luminance change parameter for the luminance change compensation using the method for estimating a luminance change parameter of a moving image according to claim 1. Video encoding program for executing a video encoding method of claim 6 in computer. Moving image encoding program recording a computer-readable record medium for executing the moving image coding method according to a computer to claim 6.

Images