JP2008141436A - Moving image encoding method, device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To avoid the distortion of a vertical boundary of a block seen when adding a weak film grain image in film grain technology. <P>SOLUTION: This moving image encoding device comprises: a noise reducing part 11 for performing noise elimination processing of an input image of an encoding target including a noise image to generate a first main image; an estimating part 12 for estimating the noise image from the input image and the first main image to generate model information; a generating part 13 for generating a block vertical boundary signal being an image component to be damaged by smoothing filter processing to a boundary between respective noise image blocks of a film grain image with reference to the model information; an adding part 14 for adding the first main image and the block vertical boundary signal to generate a second main image; an encoding part 15 for encoding the second main image to generate encoded data; and a multiplexing part 16 for multiplexing the encoded data with the model information and outputting an encoded stream. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a moving image encoding method, apparatus, and program capable of satisfactorily reproducing film grain.

  When a movie film image is encoded for the purpose of digital versatile disc (DVD) content use, there is a problem that the original atmosphere of the movie is impaired. This is because film grain in the video (film grain feeling) is lost with encoding. Film grain is also called the soul of Hollywood and is very important in DVD authoring studios.

  As a method for successfully reproducing the film grain, there is a film grain simulation method disclosed in Patent Document 1. As a representative example, H.I. There is Film Grain Technology (FGT) in H.264.

  In the encoding process in FGT, first, a main image from which film grain is removed is generated by performing film grain removal processing on an input image that is a moving image. The generated main image is subjected to compression coding processing using motion compensation, orthogonal transformation, and quantization by the coding unit. On the other hand, information for generating a film grain image from the difference between the input image and the main image by the noise estimation unit and reproducing the film grain (information for identifying the cited film grain image database, film grain strength, etc.) Is used to generate model information.

  Next, in the decoding process in FGT, a main image from which film grain is removed as a decoded image is generated by a decoding unit, and a noise image is generated from the main image using model information transmitted from the encoding unit by a noise generation unit. That is, a film grain image is generated. By adding the generated main image and the film grain image, an output image in which the film grain is reproduced is obtained.

The advantage of such an FGT is that the film grain is separated and modeled, so that the film grain is not lost by the encoding process and the original atmosphere of the movie can be reproduced.
US Patent Application Publication No. 2006/0082649

  According to the studies by the inventors, when a weak film grain image is added to the decoded image in the FGT, that is, when the value of information indicating the strength of the film grain is small among the information for reproducing the film grain. It was confirmed that the phenomenon of distortion appearing at the block vertical boundary of the output image occurred. This is a problem because it greatly deteriorates the quality of the output image. This is considered to be caused by applying a smoothing filter process called deblocking filter to the vertical edges of blocks in order to make the seams of film grain images inconspicuous in FGT film grain image generation processing. It is.

  An object of the present invention is to avoid the distortion of the vertical boundary of a block seen when adding a weak film grain image.

According to an aspect of the present invention, a step of generating a first main image by performing noise removal processing on an input image to be encoded including a noise image, and estimating the noise image from the input image and the first main image Generating model information corresponding to the noise image, and referring to the model information, an image component that is impaired by smooth filter processing on a boundary between each noise image block of the film grain image for film grain reproduction Generating a second main image by adding the first main image and the image component, encoding the second main image to generate encoded data,
And a step of multiplexing the model information with the encoded data and outputting an encoded stream.

  According to a more specific aspect of the present invention, a step of generating a first main image by performing a noise removal process on an input image to be encoded including a first noise image, and from the input image and the first main image Estimating the first noise image to generate model information including a database ID and an intensity modulation coefficient corresponding to the first noise image; and a plurality of pixels in the first main image with reference to the model information Selecting a noise database storing a noise image block according to a database ID associated with the pixel average value of the image block for each image block including: cutting out a noise image block from the noise database; and the model information Selecting an intensity modulation coefficient associated with the pixel average value with reference to Modulating the noise image block by the intensity modulation coefficient to generate a modulated noise image; and applying a smoothing filter process to a boundary between the noise image blocks of the modulated noise image to generate a smoothed noise image; Subtracting the smoothed noise image from the modulated noise image to generate a difference image; adding the first main image and the difference image to generate a second main image; and the second main image A moving picture encoding method is provided, which includes the steps of: generating encoded data by encoding the model information; and outputting the encoded stream by multiplexing the model information with the encoded data.

  Here, the step of generating the second main image may include adding the difference image to the first main image after multiplying the difference image by a coefficient.

  In the step of generating the second main image, the difference image may be added to the first main image after adding the offset to the difference image.

  According to another aspect of the present invention, a step of generating a first main image by performing a noise removal process on an input image to be encoded including a noise image, and the noise image from the input image and the first main image. A periodic pattern is generated by generating model information having a plurality of parameters corresponding to the noise image by estimation and smoothing filter processing on a boundary between each noise image block of the film grain image for film grain reproduction. The step of storing the specific parameter range, the step of determining whether or not the value of the parameter is within the specific parameter range, and the step of determining determine that the value of the parameter is within the specific parameter range. The parameter value is corrected to a value outside the specified parameter range. And a step of generating encoded data by encoding the main image, and a step of multiplexing the model information that has undergone the correcting step and outputting an encoded stream to the encoded data. An image encoding method is provided.

  Furthermore, according to another aspect of the present invention, there is provided a moving image encoding apparatus that implements the above-described moving image encoding processing by hardware, or a program for causing a computer to perform the above-described moving image encoding processing.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to avoid the distortion of the vertical boundary of the block seen when adding a weak film grain image in FGT.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. In the following description, the image signal is simply referred to as an image.

(First embodiment)
As shown in FIG. 1, a moving image encoding apparatus according to an embodiment of the present invention includes a noise reduction unit 11, a noise estimation unit 12, a block vertical boundary signal generation unit 13, an addition unit 14, an encoding unit (encoder). ) 15 and a multiplexing unit (multiplexer) 16. On the other hand, a moving picture decoding apparatus corresponding to the moving picture encoding apparatus of FIG. 1 includes a separation unit (demultiplexer) 21, a decoding unit (decoder) 22, a noise generation unit 23, and an addition unit 24 as shown in FIG. Have

  In the moving image encoding apparatus in FIG. 1, noise reduction processing is performed by the noise reduction unit 11 on the input image 101 that is a moving image input to the input terminal 10, and a main image 102 (first main image) is obtained. Generated. The noise estimation unit 12 models noise (film grain) included in the input image 101 from the difference between the input image 101 and the main image 102 and outputs model information 103. Specifically, the model information 103 is, for example, information for constructing an FGT table described later, and is also called a film grain characteristics SEI message.

  The block vertical boundary signal generation unit 13 refers to the model information 103 output from the noise estimation unit 12, and the image component that is damaged at the block vertical boundary by the FGT smoothing filter processing from the main image 102 (this image component is added) This signal is added to the block vertical boundary of the main image 102 by the unit 14, and is hereinafter referred to as a block vertical boundary signal) 104. The block vertical boundary signal 104 is added to the main image 102 from the noise reduction unit 11 in the adding unit 14 to generate a main image 105 (second main image) that can compensate for distortion of the block vertical boundary.

  The main image 105 output from the adder 14 is subjected to compression coding using motion compensation, orthogonal transformation such as DCT, and quantization by an encoding unit 15 (for example, an H.264 encoder). Is generated. The encoded data 106 is multiplexed together with the model information 103 output from the noise estimation unit 13 by the multiplexing unit 16 and output as an encoded bit stream 107. The encoded bit stream 107 is recorded on a storage medium such as a DVD or HD DVD (not shown), or sent to a transmission medium such as the Internet.

  On the other hand, in the moving picture decoding apparatus in FIG. 2, an encoded bit stream 201 (basically, an encoded stream 107 and an image reproduced from the above storage medium or sent via a transmission medium to the input terminal 20 is used. The same) is entered. In the separation unit 21, the encoded data 202 and the model information 203 are separated from the encoded bitstream 201. In the decoding unit 22, a main image (decoded image) 204 is generated by decoding the encoded data 202. The noise generation unit 108 generates a noise image (film grain image) 205 from the main image 204 according to the model information 203. Finally, by adding the main image 204 and the noise image 205 in the addition unit 24, an output image 206 of a moving image in which film grain is reproduced is obtained.

  The processing in the noise generation unit 23 includes smoothing filter processing for the block vertical boundary for the purpose of making the noise image seam inconspicuous as in Patent Document 1, and this generates distortion in the block vertical boundary. Cause. FIG. 3 is a diagram showing the situation. In the output image (result image) obtained by adding the noise image (film grain image) to the main image (decoded image), the film grain is reproduced. A vertical line, which is a distortion of the vertical boundary, is entered, resulting in a deterioration in image quality. Although different from the FGT standard, when noise generation processing is performed without performing such smoothing filter processing, it is confirmed that distortion to the block vertical boundary of the output image does not appear and a normal image can be obtained. It was.

  Therefore, in the moving picture encoding apparatus shown in FIG. 1, an image component (this is called a block vertical boundary signal) 104 that is damaged by the smoothing filter processing in the noise generating unit 23 of the moving picture decoding apparatus in FIG. Generated by the signal generator 13, the block vertical boundary signal 104 is added to the main image 102 output from the noise reduction unit 11. As a result, the output image 206 is close to the image when the smoothing filter processing in the noise generating unit 23 is omitted, so that the output image 206 is a high quality image without the distortion of the block vertical boundary as shown in FIG.

Next, the block vertical boundary signal generation unit 13 will be described in detail.
FIG. 4 is an example of the FGT table 301 which is the model information 103 generated by the noise estimation unit 12, and here has four items of range L, range H, DB_ID, and COFF. Range L and range H represent lower and upper limits set for the average pixel value of, for example, an 8 × 8 pixel block of the main image. DB_ID is an ID indicating the noise database 302 in FIG. 5 corresponding to the range indicated by the range L and the range H. COFF is an intensity (amplitude) modulation coefficient corresponding to the range indicated by the range L and the range H. In the noise database 302, for example, as shown in FIG. 5, noise image block data is stored at positions ID1, ID2,.

  Here, in the moving picture decoding apparatus shown in FIG. 2, the noise image 205 is generated by the noise generation unit 23 as follows. First, the noise image block is cut out from the noise database according to DB_ID corresponding to the range indicated by the range L and the range H in which the average pixel value of the pixel block of the main image 204 output from the decoding unit 22 is entered. The extracted noise image block is modulated by being multiplied by the intensity modulation coefficient COFF corresponding to the average pixel value of the pixel block of the main image 204, thereby generating a modulated noise image. A smoothing filter process is performed on the vertical boundary between noise image blocks of the generated modulated noise image, so that a noise image 205 to be added to the main image 204 is generated.

  The block vertical boundary signal generation unit 13 shown in FIG. 1 pays attention to the above-described processing of the noise generation unit 23 in the moving picture decoding apparatus of FIG. 2, for example, according to the procedure shown in the flowchart of FIG. 104 is generated.

  First, for example, an average pixel value of an 8 × 8 pixel block of the main image 102 (image after noise removal output from the noise reduction unit 11) is calculated (step S101). By referring to the FGT table 301 using the average pixel value as a key, DB_ID that is an ID of the noise database 302 corresponding to the average pixel value is determined (step S102). With reference to the noise database 302 using the DB_ID determined in step S102 as a key, a noise image block corresponding to the DB_ID is cut out (step S103).

  Next, as in step S102, the intensity modulation coefficient COFF corresponding to the average pixel value is determined by referring to the FGT table 301 based on the average pixel value (step S104). The noise image block cut out in step S103 is intensity-modulated according to COFF determined in step S104 (step S105).

  For example, if the average pixel value calculated in step S101 is “30”, the DB_ID corresponding to the average pixel value “30” is “2” from FIG. For this reason, DB_ID is determined to be ID2 in step S102, and a noise image block corresponding to ID2 is cut out in step S103. On the other hand, the intensity modulation coefficient COFF corresponding to the average pixel value “30” is determined to be 0.5 from FIG. 4, and 0.5 is multiplied to the noise image block cut out in step S103.

  By performing the processing of steps S101 to S105 for all image blocks of the main image 102, in step S105, a modulated noise image including noise image blocks corresponding to all image blocks of the main image 102 and subjected to intensity modulation. Is generated.

  Next, a smoothing noise image is generated by performing smoothing filter processing on the vertical boundary of the noise image block of the modulated noise image generated in step S105 (step S106). A difference image obtained by subtracting the smoothed noise image generated in step S106 from the modulated noise image generated in step S105 is generated as the block vertical boundary signal 104 (step S107).

  As described above, in the present embodiment, the block vertical boundary signal 104 generated by the block vertical boundary signal generation unit 13 in the main image 102 after noise removal output from the noise reduction unit 11 in the moving image encoding device of FIG. Then, the encoding unit 15 performs encoding. The block vertical boundary signal 104 is an image component that is impaired by the smoothing filter processing in the noise generation unit 23 of the video decoding device as described above, and is added to the main image 102 in advance before encoding.

  Therefore, in the moving picture decoding apparatus, the noise image 205 from the noise generation unit 23 is added to the main image 204 from the decoding unit 22 to reproduce the film grain, but the smoothing filter in the noise generation unit 23 is reproduced. A high-quality output image 107 without distortion of the block vertical boundary due to the processing can be obtained.

  Next, a modification of the first embodiment will be described. FIG. 7 shows a processing procedure of the block vertical boundary signal generation unit 13 according to the modification, and the processing of step S108 is added after step S107 of FIG. In step S108, the block vertical boundary signal is obtained by performing intensity modulation by multiplying the difference image obtained by subtracting the smoothed noise image from the modulated noise image in step S107 by a coefficient that takes into account deterioration due to encoding distortion. 104 is generated.

  FIG. 8 shows a processing procedure of the block vertical boundary signal generation unit 13 according to another modification, and the processing of step S109 is added after step S of FIG. In step S109, the block vertical boundary signal 104 is generated by adding an offset to the difference image obtained by subtracting the smoothed noise image from the modulation noise image in step S107.

  In the noise generating unit 23 in the moving picture decoding apparatus in FIG. 2, there is a high probability that the distortion of the block vertical boundary appears on the negative side due to rounding of the smoothing filter. Therefore, in step S109, for example, a block vertical boundary signal 104 that can more accurately compensate for block vertical boundary distortion is generated by adding a positive offset to the difference image.

  Note that the processing in step S108 in FIG. 7 and step S109 in FIG. 8 may be combined, thereby making it possible to more accurately compensate for distortion at the block vertical boundary.

(Second Embodiment)
FIG. 9 shows a video encoding apparatus according to the second embodiment of the present invention. In this embodiment, the block vertical boundary signal generation unit 13 and the addition unit 14 are removed from FIG. 1, and a parameter range determination unit 17, a specific parameter range storage unit 18, and a parameter correction unit 19 are added instead.

  The model information 103 output from the noise estimation unit 12 is input to the multiplexing unit 16 via the parameter correction unit 19. The parameter range determination unit 17 determines whether or not the model information 103 is within a predetermined range (referred to as a specific parameter range) stored in the specific parameter range storage unit 18.

  Here, the specific parameter range refers to the parameter range of the model information 103 in which the distortion of the periodic pattern (in other words, the distortion of the block vertical boundary), which is an adverse effect of FGT, is observed. FIG. 10 is a diagram illustrating an example of such a specific parameter range. In this example, in the model information 103 (FGT table) exemplified above, DB_ID is taken on the horizontal axis, and COFF is taken on the vertical axis. The range indicated by diagonal lines is the specific parameter range.

  If the input model information 103 is outside the specific parameter range, the parameter correction unit 19 outputs the model information 103 as model information 108 as it is. On the other hand, when the model information 103 is within the specific parameter range, the parameter correction unit 19 corrects the model information 103 to be out of the specific parameter and outputs the model information 108.

  As described above, according to the present embodiment, by not using the values of the range of the model information parameters that cause the distortion of the periodic pattern in the output image 206 of the video decoding device, the block vertical boundary Thus, a high-quality output image 206 in which no periodic pattern distortion such as the above-described distortion occurs is obtained.

  The moving image encoding processing based on the embodiment of the present invention described above can be realized by hardware, but can also be executed by software using a computer such as a personal computer. Therefore, according to the present invention, it is possible to provide a program for causing a computer to perform the above-described moving image encoding process or a computer-readable storage medium storing the program.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, the processing order of each step shown in the flowcharts of FIGS. 6 to 8 can be changed as appropriate. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

1 is a block diagram showing a video encoding device according to a first embodiment. 1 is a block diagram showing a moving picture decoding apparatus according to a first embodiment. Illustration for explaining image quality degradation due to side effects of FGT The figure which shows an example of the FGT table which is model information Figure showing an example of a noise database Flow chart showing processing procedure of block vertical boundary signal generation unit The flowchart which shows the other processing procedure of a block vertical boundary signal generation part The flowchart which shows another processing procedure of a block vertical boundary signal generation part. The block diagram which shows the moving image encoder which concerns on 2nd Embodiment. The figure which shows the example of the specific parameter range in 2nd Embodiment

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Moving image input terminal 11 ... Noise reduction part 12 ... Noise estimation part 13 ... Block vertical boundary signal generation part 14 ... Adder 15 ... Encoding part 16 ... Multiplexer 17... Parameter range discriminator 18... Specific parameter range storage unit 19... Parameter modifier 20... Encoded stream input terminal 21. ... Noise generator 24 ... Adder 101 ... Input image 102 ... First main image 103 ... Model information 104 ... Block vertical boundary signal 105 ... Second main image 106 .. Encoded data 107 ... Encoded bit stream 108 ... Model information 201 ... Encoded stream 202 ... Encoded data 203 ... Model information 204 The main image 205 ... noise image 206 ... output image

Claims (16)

Performing a noise removal process on an input image to be encoded including a noise image to generate a first main image;
Estimating the noise image from the input image and the first main image to generate model information corresponding to the noise image;
Referring to the model information, generating an image component that is impaired by smoothing filter processing on a boundary between each noise image block of a film grain image for film grain reproduction; and
Adding the first main image and the image component to generate a second main image;
Encoding the second main image to generate encoded data;
And a step of multiplexing the model information with the encoded data and outputting an encoded stream. Performing a noise removal process on an input image to be encoded including a first noise image to generate a first main image;
Estimating the first noise image from the input image and the first main image to generate model information including a database ID and an intensity modulation coefficient corresponding to the first noise image;
Selecting a noise database storing a noise image block for each image block including a plurality of pixels in the first main image according to a database ID associated with a pixel average value of the image block with reference to the model information When,
Cutting out a noise image block from the noise database;
Selecting an intensity modulation coefficient associated with the pixel average value with reference to the model information;
Modulating the noise image block with the intensity modulation coefficient to generate a modulated noise image;
Applying a smoothing filter process to a boundary between each noise image block of the modulated noise image to generate a smoothed noise image;
Subtracting the smoothed noise image from the modulated noise image to generate a difference image;
Adding the first main image and the difference image to generate a second main image;
Encoding the second main image to generate encoded data;
And a step of multiplexing the model information with the encoded data and outputting an encoded stream.   The moving image encoding method according to claim 2, wherein the step of generating the second main image is configured to add the difference image to the first main image after multiplying the difference image by a coefficient.   The moving image encoding method according to claim 2, wherein the step of generating the second main image is configured to add the difference image to the difference image and add the offset to the first main image. . Performing a noise removal process on an input image to be encoded including a noise image to generate a first main image;
Estimating the noise image from the input image and the first main image to generate model information having a plurality of parameters corresponding to the noise image;
Storing a specific parameter range in which a periodic pattern is generated by a smoothing filter process for a boundary between each noise image block of a film grain image for film grain reproduction;
Determining whether the value of the parameter is within the specific parameter range; and
Correcting the parameter value to a value outside the specific parameter range when the parameter value is determined to be within the specific parameter range by the determining step;
Encoding the main image to generate encoded data;
A method of multiplexing the model information that has undergone the correcting step and outputting an encoded stream to the encoded data.   The moving picture encoding method according to claim 5, wherein the model information includes a database ID and an intensity modulation coefficient corresponding to the noise image as the parameters. A noise reduction unit that generates a first main image by performing noise removal processing on an input image to be encoded including a noise image;
An estimation unit that estimates the noise image from the input image and the first main image and generates model information corresponding to the noise image;
A generation unit that refers to the model information and generates an image component that is impaired by smoothing filter processing on a boundary between each noise image block of a film grain image for film grain reproduction;
An adder for adding the first main image and the image component to generate a second main image;
An encoding unit that encodes the second main image to generate encoded data;
And a multiplexing unit that multiplexes the model information with the encoded data and outputs an encoded stream. A noise reduction unit that generates a first main image by performing a noise removal process on an input image to be encoded including a first noise image;
An estimation unit that estimates the first noise image from the input image and the first main image and generates model information including a database ID and an intensity modulation coefficient corresponding to the first noise image;
Means for selecting a noise database storing a noise image block for each image block including a plurality of pixels in the first main image according to a database ID associated with a pixel average value of the image block with reference to the model information When,
Means for extracting a noise image block from the noise database;
Means for selecting an intensity modulation coefficient associated with the pixel average value with reference to the model information;
Means for modulating the noise image block with the intensity modulation coefficient to generate a modulated noise image;
Means for applying a smoothing filter process to a boundary between each noise image block of the modulated noise image to generate a smoothed noise image;
A generating unit that subtracts the smoothed noise image from the modulated noise image to generate a difference image;
An adder for adding the first main image and the difference image to generate a second main image;
An encoding unit that encodes the second main image to generate encoded data;
And a multiplexing unit that multiplexes the model information with the encoded data and outputs an encoded stream.   The moving image encoding apparatus according to claim 8, wherein the generation unit is configured to add the difference image to the first main image after multiplying the difference image by a coefficient.   The moving image encoding apparatus according to claim 8, wherein the generation unit is configured to add the difference image to the first main image after adding the offset to the difference image. A generation unit that performs noise removal processing on an input image to be encoded including a noise image to generate a first main image;
A generation unit configured to generate model information having a plurality of parameters corresponding to the noise image by estimating the noise image from the input image and the first main image;
A storage unit for storing a specific parameter range in which a periodic pattern is generated by smoothing filter processing on a boundary between each noise image block of a film grain image for film grain reproduction;
A determination unit for determining whether or not the value of the parameter is within the specific parameter range;
A correction unit that corrects the value of the parameter to a value outside the specific parameter range when the value of the parameter is determined to be within the specific parameter range by the determining step;
An encoding unit that encodes the main image to generate encoded data;
And a multiplexing unit that multiplexes the encoded data with the model information that has passed through the correction unit and outputs an encoded stream. A process of generating a first main image by performing a noise removal process on an input image to be encoded including a noise image;
Processing for estimating the noise image from the input image and the first main image and generating model information corresponding to the noise image;
With reference to the model information, a process of generating an image component that is damaged by a smoothing filter process on a boundary between each noise image block of a film grain image for film grain reproduction;
A process of adding the first main image and the image component to generate a second main image;
A process of generating the encoded data by encoding the second main image;
A program for causing a computer to perform a moving image encoding process including: a process of multiplexing the model information with the encoded data and outputting an encoded stream. A process of generating a first main image by performing a noise removal process on an input image to be encoded including a first noise image;
Processing for estimating the first noise image from the input image and the first main image and generating model information including a database ID and an intensity modulation coefficient corresponding to the first noise image;
Processing for selecting a noise database storing a noise image block for each image block including a plurality of pixels in the first main image according to a database ID associated with a pixel average value of the image block with reference to the model information When,
A process of cutting out a noise image block from the noise database;
Selecting an intensity modulation coefficient associated with the pixel average value with reference to the model information;
Processing to generate a modulated noise image by modulating the noise image block with the intensity modulation coefficient;
Processing to generate a smoothed noise image by applying a smoothing filter process to a boundary between each noise image block of the modulated noise image;
A process of subtracting the smoothed noise image from the modulated noise image to generate a difference image;
A process of generating the second main image by adding the first main image and the difference image;
A process of generating the encoded data by encoding the second main image;
A program for causing a computer to perform a moving image encoding process including: a process of multiplexing the model information with the encoded data and outputting an encoded stream.   The program according to claim 13, wherein the process of generating the second main image is a process of adding the difference image to the first main image after multiplying the difference image by a coefficient.   The program according to claim 13, wherein the process of generating the second main image is a process of adding the difference image to the first main image after adding an offset to the difference image. A process of generating a first main image by performing a noise removal process on an input image to be encoded including a noise image;
Processing to generate model information having a plurality of parameters corresponding to the noise image by estimating the noise image from the input image and the first main image;
A process for storing a specific parameter range in which a periodic pattern is generated by a smoothing filter process on a boundary between each noise image block of a film grain image for film grain reproduction;
Processing for determining whether or not the value of the parameter is within the specific parameter range;
A process for correcting the parameter value to a value outside the specific parameter range when the parameter value is determined to be within the specific parameter range by the determining process;
A process of encoding the main image to generate encoded data;
A program for causing a computer to perform a moving image encoding process including: a process of multiplexing model information that has undergone the correction process to the encoded data and outputting an encoded stream.

Images