JP2012235187A - Noise detection apparatus, reproduction apparatus, and noise detection program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide means for accurate and automatic detection of characteristic noise contained in an image.SOLUTION: A noise detection apparatus comprises: an acquisition part which acquires information on a target image to be processed; a calculation part which calculates a feature amount relating to a noise to be detected, on the basis of the target image; and a determination part which determines whether the target image contains the noise to be detected, according to the feature amount and at least one of the shape of the noise to be detected, information on brightness, and information on colors.

Description

  The present invention relates to a noise detection device, a playback device, and a noise detection program.

  In an image, it is known that inherent noise is generated due to a specific factor. For example, as a factor of inherent noise generated in a moving image, there are a mechanical factor when capturing a moving image, a mechanical factor when copying a moving image between a plurality of devices, a factor in software, and the like. In any case, such noise must be detected visually, and there is a problem that time and labor are required for detection. In addition, the noise described above often occurs instantaneously (for example, only a few frames), and may be overlooked unless carefully checked.

  Therefore, it is considered to automate such noise detection. For example, the invention of Patent Document 1 discloses a technique for automatically detecting an abnormality in a video information component from a video signal itself.

Japanese Patent Laid-Open No. 9-46733

  However, the invention of Patent Document 1 is merely to detect a video freeze state and a scene change, and cannot perform fine detection such as the above-described noise detection.

  An object of the present invention is to provide means for accurately and automatically detecting inherent noise included in an image.

  A noise detection apparatus according to one aspect includes an acquisition unit that acquires information on a target image to be processed, a calculation unit that calculates a feature amount related to noise to be detected based on the target image, and a detection target A determination unit configured to determine whether or not the detection target noise is included in the target image based on at least one of noise shape, information on brightness, information on color, and the feature amount; Is provided.

  The calculation unit calculates at least one of luminance and brightness as the feature amount, and the determination unit adds the detection target to the target image based on the feature amount and the information on the brightness. It may be determined whether or not the noise is included.

  Further, the calculation unit calculates at least one of an RGB value, saturation, hue, and color difference as the feature amount, and the determination unit converts the feature amount and information about the color. Based on this, it may be determined whether or not the target image includes noise to be detected.

  In addition, the calculation unit calculates at least the saturation as the feature amount, and the determination unit includes noise of the detection target in the target image based on the saturation and the information on the color. It may be determined whether or not.

  In addition, the calculation unit calculates at least the hue as the feature amount, and the determination unit includes noise of the detection target in the target image based on the hue and information on the color. You may determine whether or not.

  The noise to be detected may have a block shape.

  The noise to be detected may have a square shape with 8 pixels on one side, or a rectangular shape with 8 pixels on one side and 16 pixels on the other side.

  The acquisition unit acquires information on a moving image including a plurality of frames as the target image, the calculation unit calculates the feature amount for each of the plurality of frames, and the determination unit includes the determination unit It may be determined whether or not the detection target noise is included for each of a plurality of frames.

  The reproduction apparatus according to one aspect includes a recording unit that records an image, a reproduction unit that reproduces the image, and any of the noise detection devices described above, and the acquisition unit records information on the target image. Get from the department.

  The noise detection program according to one aspect includes an acquisition process for acquiring information of a target image to be processed, a calculation process for calculating a feature amount related to noise to be detected based on the target image, and a detection target A determination process for determining whether or not the detection target noise is included in the target image, based on at least one of noise shape, information on brightness, information on color, and the feature amount; Is executed on the computer.

  According to the present invention, it is possible to provide a means for accurately and automatically detecting inherent noise included in an image.

Block diagram showing a configuration example of a noise detection device Diagram explaining the noise to be detected Flow chart showing an example of operation of the noise detection device when detecting block noise α Diagram explaining detection of block noise α

  FIG. 1 is a block diagram illustrating a configuration example of a noise detection device according to an embodiment. In the noise detection apparatus, a noise detection program for detecting noise is installed in advance for an image to be processed (target image). Hereinafter, a case where noise in a moving image is detected as a target image will be described as an example. Note that the same processing may be performed when the target image is a still image.

  Further, any target moving image may be used. For example, it may be generated by an imaging device capable of generating a digital image, may be converted from analog moving image image data to digital data, or recorded on a video tape or the like. A moving image may be converted into digital data. Further, it may be a moving image created by a computer or the like.

  In the following, an example in which all the frames of the target moving image are processed as noise detection targets will be described. However, only a part of the frames of the target moving image may be processed as noise detection targets. In this case, the noise detection target frame may be specified based on a user operation, or may be automatically specified based on information on the target moving image.

  The noise detection device 11 shown in FIG. 1 includes a noise detection unit 13, a CPU 14, an alarm output unit 15, an input / output I / F 16, and a bus 17. The CPU 14 is connected to the input / output I / F 16. The output of the noise detection unit 13 is connected to the CPU 14 via the bus 17. Further, the alarm output unit 15 is controlled by the CPU 14.

  The image data of the target moving image input to the noise detection device 11 is input to the noise detection unit 13. The image data of the target image input to the noise detection unit 13 is YCbCr format image data. Furthermore, an input device (such as a keyboard and pointing device) (not shown), a monitor, a network cable for LAN connection, and the like are connected to the noise detection device 11 via an input / output I / F 16.

  The noise detection device 11 includes a storage device (not shown) (for example, a storage medium such as a hard disk or a nonvolatile semiconductor memory). This storage device stores an image processing program and various data necessary for executing the program.

  The CPU 14 is a processor that comprehensively controls each unit of the noise detection device 11. The alarm output unit 15 includes a notification member such as a speaker, a monitor, and a lamp, and notifies the user of the detection result by the noise detection unit 13.

  The noise detection apparatus 11 having the configuration described above detects predetermined inherent noise. Any type of noise may be detected. In the following, three types of noise (block noise α to block noise γ) will be described as an example.

  FIG. 2 is a diagram illustrating the three types of noise described above.

  The block noise α is rectangular noise having a large luminance difference compared to the surroundings. The block noise β is a rectangular noise with relatively low saturation compared with the surrounding area. Therefore, the block noise β is substantially colorless block noise. The block noise γ is a rectangular noise having a yellow to green color.

  Further, the three types of noise described above are all noise generated in a block shape as shown in FIG. In the following, as an example, square-shaped noise (block noise αa to block noise γa) shown in the enlarged view E1 and rectangular noise (block noise αb to block noise γb) shown in the enlarged view E2 are given as examples. explain. In addition, in the following, as an example, the square-shaped noise shown in the enlarged view E1 has a size of 8 pixels on one side, and the rectangular-shaped noise shown in the enlarged view E2 has 8 pixels on one side and 16 pixels on the other side An example will be described.

  The 8 pixels described above are MPEG2 and H.264. The size corresponds to a processing unit in the compression processing such as H.264. In addition, the three types of noise described above often occur suddenly only in frames of about 1 to 3 frames.

  As shown in FIG. 1, the noise detection unit 13 includes a color space conversion unit 21, a line memory unit 22, a block noise α detection unit 23 that detects block noise α, a block noise β detection unit 24 that detects block noise β, A block noise γ detector 25 for detecting the block noise γ is included.

  The color space conversion unit 21 receives the image data in the YCbCr format described above. The color space conversion unit performs a known color space conversion process on the image data in the YCbCr format, and generates image data in the HSV format. The color space conversion unit 21 corresponds to the Y signal corresponding to the Y component (luminance component) in the YCbCr format image data and the S component (saturation component) in the HSV image data. The S signal and the H signal corresponding to the H component (hue component) are output to the line memory unit 22.

  The line memory unit 22 outputs the Y signal described above to the block noise α detection unit 23 by two lines. Further, the line memory unit 22 outputs the above-described S signal for each line to the block noise β detection unit 24. Further, the line memory unit 22 outputs the above-described H signal for two lines to the block noise γ detection unit 25.

  Next, details of noise detection in each unit will be described.

  First, an example of the operation of detecting the block noise α in the block noise α detecting unit 23 will be described with reference to the flowchart of FIG. The block noise α detection unit 23 detects the block noise α by performing a combination of edge detection and luminance level detection according to the characteristics of the block noise α (αa and αb) that is the detection target.

  3 is started when each unit executes a noise detection program in response to a program execution instruction from the user.

(Step S101)
The noise detection unit 13 acquires image data of one frame of the target moving image specified by the user. The acquired image data of one frame is input to the block noise α detection unit 23 via the line memory unit 22 as shown in FIG. At this time, the Y signal corresponding to the Y component (luminance component) is input to the block noise α detection unit 23 by two lines.

(Step S102)
The block noise α detector 23 determines whether or not an upper edge exists. If the block noise α detection unit 23 determines that there is an upper edge, the block noise α detection unit determines that there is a block noise candidate and proceeds to step S103. If the block noise α detection unit 23 determines that there is no upper edge, the block noise α detection unit 23 determines that there is no block noise candidate. Proceed to S114.

  FIG. 4A shows an upper edge ET as an example of the upper edge. The upper edge ET corresponds to the upper side of the block noise candidate described above. The block noise α detection unit 23 compares the Y signal values of the adjacent pixels above and below based on the above-described two lines of Y signals, and if the difference between the Y signals is greater than a predetermined threshold value, Further, the Y signal values of pixels adjacent in the vertical direction are compared in the adjacent pixels on the right side. The same processing is repeated, and when the difference between the Y signals of pixels adjacent in the vertical direction over a pixel having a predetermined length (for example, 8 pixels) is larger than a predetermined threshold, the block noise α detection unit 23 It is determined that the upper edge exists. Hereinafter, the pixel position at the left end of the upper edge (the upper left end of the block noise candidate) is referred to as a position PE.

(Step S103)
If the block noise α detector 23 determines that the upper edge is present in step S102, the block noise α detector 23 sets the flag Fu = 1 indicating the presence of the upper edge and proceeds to step S104.

(Step S104)
The block noise α detector 23 determines whether or not a left edge exists. If the block noise α detector 23 determines that the left edge exists, the block noise candidate is determined to be correct on the left side, and the process proceeds to step S105. If it is determined that the left edge does not exist, the block noise α detector 23 determines that the left edge exists. Considering that the block candidate is an error, the process proceeds to step S112 described later.

  FIG. 4A shows a left edge EL as an example of the left edge. The block noise α detection unit 23 selects the lowest line among the currently read lines (or the (n + 1) th line if the read lines are the nth and (n + 1) th lines). Among the signals, the pixel position PE (or the pixel adjacent to the lower side of the pixel position PE) is used as a target pixel, and the Y signal values of pixels adjacent to the left and right are compared, and the difference in Y signal is greater than a predetermined threshold value. In the case, it is determined that the left edge exists.

(Step S105)
The block noise α detection unit 23 determines whether or not the in-block level satisfies a predetermined condition. If the block noise α detection unit 23 determines that the in-block level satisfies a predetermined condition, the block noise α detection unit 23 proceeds to step S106. If the block noise α detection unit 23 determines that the in-block level does not satisfy the predetermined condition, the block noise α detection unit 23 proceeds to step S112 described later.

  The intra-block level indicates the level of the Y signal of the pixel in the intra-block area Ei shown in FIG. 4B. As shown in FIG. 4B, the in-block area Ei corresponds to a square area of 8 pixels square with the position PE as the upper left corner. The block noise α detection unit 23 sequentially moves the target pixel to the right in the line where the left edge is detected in step S104, and determines whether the level of the Y signal in each pixel is higher than a predetermined level. . If the block candidate determined to be present in step S102 is correct, the Y signal of the pixel in the block inner area Ei described above should always be higher than a predetermined level.

  The predetermined level used for the determination can be determined in advance according to the characteristics of the block noise α. When the level of the Y signal of all pixels existing in the block area Ei is higher than a predetermined level in the line where the left edge is detected in step S104, the block noise α detection unit 23 determines the level in the block. If the Y signal level is lower than the predetermined level even in one pixel, it is determined that the in-block level does not satisfy the predetermined condition.

  In step S105, when determining whether the in-block level satisfies a predetermined condition, the in-block area Ei shown in FIG. A rectangular area of 16 pixels in the vertical direction (area corresponding to the block noise αb) may be used.

(Step S106)
The block noise α detector 23 determines whether or not the right edge (1) exists. The right edge (1) is an edge corresponding to the right side of the block noise αa which is a square-shaped noise among the block noise α described with reference to FIG. The block noise α detection unit 23 determines whether or not the right edge (1) exists, similarly to the left edge described in step S104. If the block noise α detection unit 23 determines that the right edge (1) exists, the block noise candidate is regarded as correct on the right side, and the process proceeds to step S108 described later. If the right edge (1) does not exist. If it determines, it will progress to step S107. FIG. 4A shows a right edge ER (1) as an example of the right edge (1).

(Step S107)
The block noise α detector 23 determines whether or not the right edge (2) exists. The right edge (2) is an edge corresponding to the right side of the block noise αb that is rectangular noise among the block noise α described with reference to FIG. The block noise α detection unit 23 determines whether or not the right edge (2) exists, similarly to the left edge described in step S104. If the block noise α detection unit 23 determines that the right edge (2) exists, the block noise candidate is determined to be correct on the right side, and the process proceeds to step S108. If the block noise α detection unit 23 determines that the right edge (2) does not exist. The block candidate determined to exist in step S102 is regarded as an error, and the process proceeds to step S112 described later.

  FIG. 4A shows a right edge ER (2) as an example of the right edge (2). The block noise α detector 23 can detect the two types of noise (block noise αa and block noise αb) described with reference to FIG. 2 based on the determinations in step S106 and step S107.

(Step S108)
If the block noise α detection unit 23 determines that the left edge exists in step S104 and determines that the right edge exists in step S106 or step S107, the flag Fs = 1 indicating that there is an edge on the left and right. The process proceeds to step S109.

(Step S109)
The block noise α detection unit 23 determines whether or not the processing from step S104 to step S108 has been completed for L lines. If the block noise α detection unit 23 determines that the processing has been completed for the L lines, the process proceeds to step S111 described later. If the block noise α detection unit 23 determines that the processing is not completed for the L lines, the process proceeds to step S110.

  Here, L is a value (L = 8 here) corresponding to the length (number of pixels) in the vertical direction of the two types of noise (block noise αb and block noise αb) described with reference to FIG. is there. When the block noise α detector 23 determines that the left edge is present in step S104 and determines that the right edge is present in step S106 or step S107, the block noise α detector 23 increases the value of a counter (not shown). When the value of L becomes L, it is determined that the processing from step S104 to step S108 has been completed for L lines.

  Note that the determination may be made by giving an error in the upper and lower sides to the value of L described above. That is, the block noise α to be detected is detected on the premise that the block noise α has a size of 8 pixels in the vertical direction, but actually, there may be some errors (for example, 6 to 10 pixels). is there. Therefore, by providing a width to the above-described value of L, it is possible to similarly detect noise having a slight error. Similarly, the right edge described in step S106 and step S107 may also be determined with an error on the left and right.

(Step S110)
The block noise α detector 23 sets the next line as a detection target and returns to step S104.

(Step S111)
The block noise α detector 23 determines whether or not a lower edge exists. If it is determined that the lower edge exists, the block noise α detection unit 23 determines that the block noise candidate is also correct on the lower side and proceeds to step S113 described later. If it is determined that the lower edge does not exist, the block noise α detection unit 23 determines that the lower edge exists. The block candidate is regarded as an error, and the process proceeds to step S112.

  FIG. 4A shows a lower edge EB as an example of the lower edge. The lower edge EB corresponds to the lower side of the block noise candidate described above. Similarly to the upper edge described above, the block noise α detection unit 23 determines whether or not a lower edge exists based on the two lines of Y signals.

(Step S112)
When the block candidate determined to be present in step S102 is an error (any of step S104NO, step S105NO, step S107NO, or step S111NO), the block noise α detection unit 23 returns the detection target to the position PE. Return to step S102. That is, the block noise α detection unit 23 considers that the block candidate detected in step S102 is an error, returns to the position PE that is the upper left corner of the block candidate, and performs the processing from step S102 onward for the next pixel.

(Step S113)
The block noise α detection unit 23 outputs information indicating that the block noise α is present in the image data of one frame read in step S101 to the CPU 14 via the bus 17 as the determination content. Note that the block noise α detection unit 23 includes information indicating whether the existing block noise α is the block noise αa or the block noise αb in the information according to the determination results of step S106 and step S107. .

(Step S114)
The CPU 14 determines whether or not the processing from step S102 to step S113 has been performed to the right end of the line (for one line). If the CPU 14 determines that the process has been performed to the right end of the line, the process proceeds to step S115. If the CPU 14 determines that the process has not been performed to the right end of the line, the process returns to step S102, and the processes subsequent to step S102 are performed on the next pixel at the position PE.

(Step S115)
The CPU 14 determines whether or not the processing from step S102 to step S114 has been performed for one frame. If the CPU 14 determines that the process is completed for one frame, the process proceeds to step S117, which will be described later. If the CPU 14 determines that the process is not completed for one frame, the process proceeds to step S116.

(Step S116)
The block noise α detector 23 sets the next line as a detection target and returns to step S102. That is, as shown in FIG. 4C, the block noise α detection unit 23 sequentially changes the portion to be detected from left to right and from top to bottom, and performs block noise α for the entire frame read out in step S101. Detection is performed.

(Step S117)
The CPU 14 determines whether or not the processing from step S102 to step S116 has been performed for all frames. If the CPU 14 determines that the process has been completed for all frames, the CPU 14 ends the series of processes. If the CPU 14 determines that the process has not been completed for all frames, the process proceeds to step S118.

(Step S118)
The noise detection unit 13 sets the next frame as a detection target and returns to step S101. That is, in step S101, image data of the next frame after the previously acquired frame is acquired.

  Next, an operation example of detection of block noise β in the block noise β detection unit 24 will be described. The detection of the block noise β is started when each unit executes the noise detection program in accordance with a program execution instruction from the user. In the following, only the parts different from the block noise α detection operation in the block noise α detection unit 23 described with reference to FIG. 3 will be described.

(Step S101)
Similarly to the detection of the block noise α, the noise detection unit 13 acquires one frame of image data of the target moving image specified by the user. However, as shown in FIG. 1, an S signal corresponding to an S component (saturation component) is input to the block noise β detection unit 24 line by line.

(Step S102)
The block noise β detection unit 24 determines whether an upper edge exists. However, the block noise β detection unit 24 compares the level of the S signal with a predetermined level at the upper edge ET among the S signals for one line input in step S101, so that the upper edge exists. It is determined whether or not. The predetermined level used for the determination can be determined in advance according to the characteristics of the block noise β. In the presence of block noise β, the S signal of the leftmost pixel of the upper edge should always be lower than a predetermined level.

  When the block noise β detection unit 24 detects a pixel in which the level of the S signal is lower than the predetermined level, the block noise β detection unit 24 further determines whether the level of the S signal is lower than the predetermined level in the adjacent pixel on the right side. To do. The same process is repeated, and when the level of the S signal is lower than the predetermined level over a pixel having a predetermined length (for example, 8 pixels), the block noise β detection unit 24 determines that the upper edge exists. judge.

(Step S103)
The block noise β detection unit 24 performs the same processing as the detection of the block noise α.

(Step S104)
The block noise β detection unit 24 determines whether there is a left edge. However, similarly to the upper edge shown in step S102, the block noise β detection unit 24 compares the level of the S signal with a predetermined level at the left edge EL to determine whether there is a left edge. Determine whether.

(Step S105)
The block noise β detector 24 determines whether the in-block level satisfies a predetermined condition. However, the block noise β detection unit 24 determines whether or not the level of the S signal in each pixel is lower than a predetermined level, and exists in the in-block area Ei on the line where the left edge is detected in step S104. When the level of the S signal of all the pixels to be performed is lower than the predetermined level, it is determined that the in-block level satisfies the predetermined condition, and the level of the S signal is higher than the predetermined level even for one pixel Is determined that the in-block level does not satisfy a predetermined condition.

(Step S106)
The block noise β detector 24 determines whether or not the right edge (1) exists. However, similarly to the left edge shown in step S104, the block noise β detection unit 24 compares the level of the S signal in each pixel with a predetermined level to determine whether the right edge (1) exists. Determine whether or not.

(Step S107)
The block noise β detection unit 24 determines whether or not the right edge (2) exists. However, similarly to the right edge (1) shown in step S106, the block noise β detection unit 24 compares the level of the S signal in each pixel with a predetermined level, so that the right edge (2) is detected. Determine if it exists.

(Step S108 to Step S110)
The block noise β detection unit 24 performs the same processing as the detection of the block noise α.

(Step S111)
The block noise β detector 24 determines whether or not a lower edge exists. However, similarly to the upper edge shown in step S102, the block noise β detection unit 24 compares the level of the S signal with a predetermined level at the lower edge EB to determine whether or not the lower edge exists. judge.

(Step S112)
The block noise β detection unit 24 performs the same processing as the detection of the block noise α.

(Step S113)
The block noise β detection unit 24 outputs information indicating that the block noise β exists in the image data of one frame read in step S101 to the CPU 14 via the bus 17 as the determination content. Note that the block noise β detection unit 24 includes information indicating whether the existing block noise β is the block noise βa or the block noise βb in the information according to the determination results of step S106 and step S107. .

(Step S114 to Step S118)
The block noise β detection unit 24 performs the same processing as the detection of the block noise α.

Note that the block noise β is detected based on the S signal corresponding to the S component (saturation component) as described above. The accuracy of the S signal depends on the accuracy of the Y signal corresponding to the Y component (luminance component). That is, when the level of the Y signal is low, the accuracy of the S signal decreases, and as a result, the detection accuracy of the block noise β based on the S signal also decreases. Therefore, the Y signal level is inspected (for example, a comparison between the Y signal level and a predetermined threshold) before the above-described series of block noise β is detected, and only when the Y signal is sufficiently reliable. The block noise β may be detected. Alternatively, when detecting a series of block noise β, the level of the Y signal may be inspected, and information indicating the reliability of the Y signal may be added to the detection result of the block noise β. An example of the operation of detecting the block noise γ in the detection unit 25 will be described. The detection of the block noise γ is started when each unit executes the noise detection program in accordance with a program execution instruction from the user. In the following, only the parts different from the block noise α detection operation in the block noise α detection unit 23 described with reference to FIG. 3 will be described.

(Step S101)
Similarly to the detection of the block noise α, the noise detection unit 13 acquires one frame of image data of the target moving image specified by the user. However, as shown in FIG. 1, an H signal corresponding to an H component (hue component) is input to the block noise γ detection unit 25 by two lines.

(Step S102)
The block noise γ detection unit 25 determines whether an upper edge exists. However, the block noise γ detection unit 25 compares the values of the H signals of the adjacent pixels above and below based on the above-described two lines of the H signal, and if the difference between the H signals is greater than a predetermined threshold value, Further, regarding the “possibility of an edge”, in the adjacent pixels on the right side, the values of the H signals of the adjacent pixels on the upper and lower sides are compared. The same processing is repeated, and when the difference between the H signals of adjacent pixels above and below a predetermined length (for example, 8 pixels) is larger than a predetermined threshold, the block noise α detection unit 23 It is determined that the upper edge exists.

(Step S103)
The block noise γ detection unit 25 performs the same process as the detection of the block noise α.

(Step S104)
The block noise γ detection unit 25 determines whether or not a left edge exists. However, similarly to the detection of the block noise β, the block noise γ detection unit 25 compares the level of the H signal with a predetermined level at the left edge EL to determine whether or not the left edge exists. judge. The predetermined level used for the determination can be determined in advance according to the characteristics of the block noise γ. When block noise γ is present, the H signal of the pixel at the left edge EL should always be a value corresponding to a range from green to yellow. The block noise γ detection unit 25 determines that the left edge exists when the level of the H signal is a level corresponding to a range from green to yellow.

(Step S105)
The block noise γ detection unit 25 determines whether the in-block level satisfies a predetermined condition. However, the block noise γ detection unit 25 determines whether or not the level of the H signal in each pixel is a level corresponding to a range from green to yellow, and in the line where the left edge is detected in step S104, the block noise γ detection unit 25 If the H signal level of all the pixels existing in the inner area Ei is a level corresponding to the range from green to yellow, it is determined that the in-block level satisfies a predetermined condition, and even one pixel is When the level of the H signal is a level corresponding to outside the range from green to yellow, it is determined that the in-block level does not satisfy a predetermined condition.

(Step S106)
The block noise γ detector 25 determines whether or not the right edge (1) exists. However, similarly to the left edge shown in step S104, the block noise γ detection unit 25 compares the level of the H signal in each pixel with a predetermined level to determine whether the right edge (1) exists. Determine whether or not.

(Step S107)
The block noise γ detection unit 25 determines whether or not the right edge (2) exists. However, similarly to the right edge (1) shown in step S106, the block noise γ detection unit 25 compares the level of the H signal in each pixel with a predetermined level so that the right edge (2) is detected. Determine if it exists.

(Step S108 to Step S110)
The block noise γ detection unit 25 performs the same process as the detection of the block noise α.

(Step S111)
The block noise γ detection unit 25 determines whether or not a lower edge exists. However, similarly to the upper edge shown in step S102, the block noise γ detector 25 determines whether or not the lower edge exists by comparing the values of the H signals of pixels adjacent vertically.

(Step S112)
The block noise γ detection unit 25 performs the same process as the detection of the block noise α.

(Step S113)
The block noise γ detection unit 25 outputs information indicating that the block noise γ exists in the image data of one frame read out in step S101 to the CPU 14 via the bus 17 as the determination content. Note that the block noise γ detection unit 25 includes information indicating whether the existing block noise γ is the block noise γa or the block noise γb in accordance with the determination results of step S106 and step S107. .

(Step S114 to Step S118)
The block noise γ detection unit 25 performs the same process as the detection of the block noise α.

As described above, the block noise γ is detected based on the H signal corresponding to the H component (hue component). The accuracy of the H signal depends on the accuracy of the Y signal corresponding to the Y component (luminance component) and the S signal corresponding to the S component (saturation component). That is, when the level of the Y signal or the S signal is low, the accuracy of the H signal is lowered, and as a result, the detection accuracy of the block noise γ based on the H signal is also lowered. Therefore, before the above-described series of block noise γ is detected, the level of the Y signal or H signal is inspected (for example, the level of the Y signal or H signal is compared with a predetermined threshold), and the Y signal or H signal is checked. However, the configuration may be such that the block noise γ is detected only when is sufficiently reliable. Alternatively, when detecting a series of block noise γ, the level of the Y signal or H signal may be inspected, and information indicating the reliability of the Y signal or H signal may be added to the detection result of the block noise γ. As described above, the noise detection apparatus according to the present embodiment acquires information about a target image to be processed, and calculates a feature amount related to noise to be detected based on the acquired target image. Then, based on at least one of the shape of noise to be detected, information on brightness, information on color, and a feature amount, it is determined whether or not noise to be detected is included in the target image.

  Therefore, according to the configuration of the present embodiment, the inherent noise included in the image can be accurately and automatically detected without performing time-consuming and laborious visual detection.

  Further, according to the present embodiment, the luminance is calculated as the feature amount, and whether or not the detection target noise is included in the target image is determined based on the luminance and the information on the brightness of the detection target noise. I do. Therefore, when the noise to be detected is inherent noise including a change in luminance, it can be accurately and automatically detected.

  In addition, according to the present embodiment, saturation and hue are calculated as feature amounts, and whether or not the detection target noise is included in the target image based on the feature amount and information regarding the color of the detection target noise. Judgment is made. Therefore, when the noise to be detected is a unique noise including a color change, it can be automatically detected accurately.

  In addition, according to the present embodiment, as the target image, moving image information including a plurality of frames is acquired, the feature amount is calculated for each of the plurality of frames, and noise to be detected is detected for each of the plurality of frames. It is determined whether or not it is included. Therefore, even when the target image is a moving image, inherent noise included in the image can be accurately and automatically detected without performing visual and time-consuming detection.

  In the above-described example, an example in which a series of processing is executed in accordance with a program execution instruction by the user has been shown, but the present invention is not limited to this example. For example, each time the noise detection device 11 reads some still image or moving image data, a series of processes may be automatically executed.

<Supplementary items of the embodiment>
(1) Each threshold value etc. which were demonstrated by the said embodiment are examples, and this invention is not limited to this example.

  (2) The method of each determination in the noise detection of the said embodiment is an example, and this invention is not limited to this example. For example, edge detection or level detection may be performed by a method other than that described in the above embodiment.

  (3) In the noise detection of the above embodiment, the Y signal corresponding to the Y component (luminance component), the S signal corresponding to the S component (saturation component), and the H signal corresponding to the H component (hue component). However, the present invention is not limited to this example. For example, instead of the Y signal, a V signal corresponding to a V component (brightness component) may be used. In addition to the Y signal, a V signal may be used. Further, instead of the S signal and the H signal, a CbCr signal corresponding to a CbCr component (color difference component) may be used. In addition to the S signal and the H signal, a CbCr signal may be used. Further, RGB signals may be used instead of the above signals, or RGB signals may be added to the above signals.

  (4) The noise (block noise α to block noise γ) shown in the above embodiment is an example, and the present invention is not limited to this example. For example, the noise (block noise α to block noise γ) shown in the above embodiment is an example of a square shape having 8 pixels on one side or a rectangular shape having 8 pixels on one side and 16 pixels on the other side. The present invention is also applied to block noise in which these block noises are enlarged or reduced (for example, a square shape having 16 pixels on one side or a rectangular shape having 4 pixels on one side and 8 pixels on the other side). The same can be applied. Such enlarged or reduced block noise often occurs when enlargement processing or reduction processing is performed in the process of editing the target image.

  Furthermore, the present invention can be similarly applied to inherent noise having characteristics other than those described in the above embodiment. For example, the present invention is similarly applied to intrinsic noises having shapes other than block shapes (for example, line shapes), intrinsic noises brighter than the surroundings, intrinsic noises with colors other than green to yellow, and the like. be able to.

  (5) A noise detection program for executing processing by the noise detection apparatus described in the above embodiment is also effective as a specific aspect of the present invention. This noise detection program may be stored in a storage medium such as a magnetic disk, or may be downloaded via the Internet or the like. In addition, an image processing apparatus, an imaging apparatus, and a reproduction apparatus (for example, a photo viewer, a digital photo frame, various printing apparatuses, and the like) including the noise detection apparatus described in the above embodiment are also effective as specific embodiments of the present invention. The playback apparatus described above may be configured to execute a series of processes when playing back an image.

  From the above detailed description, features and advantages of the embodiments will become apparent. It is intended that the scope of the claims extend to the features and advantages of the embodiments as described above without departing from the spirit and scope of the right. Further, any person having ordinary knowledge in the technical field should be able to easily come up with any improvements and modifications, and there is no intention to limit the scope of the embodiments having the invention to those described above. It is also possible to use appropriate improvements and equivalents within the scope disclosed in.

DESCRIPTION OF SYMBOLS 11 ... Noise detection apparatus, 13 ... Noise detection part, 14 ... CPU

Claims (10)

An acquisition unit for acquiring information of a target image to be processed;
Based on the target image, a calculation unit that calculates a feature amount related to noise to be detected;
Based on at least one of the shape of the noise to be detected, information on brightness, information on color, and the feature amount, whether or not the noise to be detected is included in the target image is determined. A noise detection apparatus comprising: a determination unit that performs the determination. The noise detection device according to claim 1,
The calculation unit calculates at least one of luminance and brightness as the feature amount,
The determination unit determines whether or not the detection target noise is included in the target image based on the feature amount and the information on the brightness.
The noise detection apparatus characterized by the above-mentioned. The noise detection device according to claim 1,
The calculation unit calculates at least one of RGB value, saturation, hue, and color difference as the feature amount,
The determination unit determines whether or not the detection target noise is included in the target image based on the feature amount and the information on the color.
The noise detection apparatus characterized by the above-mentioned. In the noise detection device according to claim 3,
The calculation unit calculates at least the saturation as the feature amount,
The determination unit determines whether or not the detection target noise is included in the target image based on the saturation and the information on the color.
The noise detection apparatus characterized by the above-mentioned. In the noise detection device according to claim 3,
The calculation unit calculates at least the hue as the feature amount,
The determination unit determines whether or not noise of the detection target is included in the target image based on the hue and information on the color.
The noise detection apparatus characterized by the above-mentioned. In the noise detection device according to any one of claims 1 to 5,
The noise to be detected has a block shape. The noise detection device according to claim 6,
The noise to be detected has a square shape with 8 pixels on one side or a rectangular shape with 8 pixels on one side and 16 pixels on the other side. In the noise detection device according to any one of claims 1 to 7,
The acquisition unit acquires information on a moving image including a plurality of frames as the target image,
The calculation unit calculates the feature amount for each of the plurality of frames,
The determination unit determines whether or not the noise to be detected is included for each of the plurality of frames. A recording unit for recording images;
A playback unit for playing back the image;
A noise detection device according to any one of claims 1 to 8,
The acquisition unit acquires information of the target image from the recording unit. An acquisition process for acquiring information of a target image to be processed;
Based on the target image, a calculation process for calculating a feature amount related to noise to be detected;
Based on at least one of the shape of the noise to be detected, information on brightness, information on color, and the feature amount, whether or not the noise to be detected is included in the target image is determined. A noise detection program that causes a computer to execute the judgment process.

Images