JP2007249436A - Image signal processor and processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an edge-reinforced image effectively using a display dynamic range within a range never reaching a saturated level by dynamically changing a reinforcement gain according to an image. <P>SOLUTION: This image signal processor comprises a local average value calculation part 11 calculating a local average value; a local component value calculation part 21 calculating a local component value that is a difference between input value and local average value; a local gain calculation and storage part 22 calculating a local gain that is a value obtained by dividing a difference between display dynamic range and local average value by the local component value; a reinforcement gain calculation part 23 taking the minimum value of a plurality of local gains as the reinforcement gain; and an edge reinforcement processing part 11 obtaining the edge-reinforced image by adding the local average value to a value obtained by multiplying the difference between input value and local average value by the reinforcement gain. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an image signal processing apparatus and an image signal processing method, and in particular, an image signal processing apparatus and an image signal that can obtain an edge enhanced image that effectively uses a display dynamic range by appropriately setting an enhancement gain. It relates to the processing method.

  As a process for sharpening an edge, that is, an outline of an image, a process for enhancing a high-frequency component of the image (hereinafter referred to as an edge enhancement process) is performed. In the edge enhancement processing, generally, a difference between an average level of a target pixel and a plurality of surrounding pixels and a level of the target pixel is obtained, and a value obtained by multiplying the difference by a predetermined enhancement gain (enhancement coefficient). Is added to the average level to enhance the high frequency components of the image. As a result, as shown in FIG. 10, the edge of the input signal is emphasized to obtain an image that is beautiful or easy for the viewer to feel.

  For example, even if the edge enhancement unit does not have multiple gain values, the scene of the processing target image is determined based on the shooting mode selection operation or subject magnification or luminance distribution information for the purpose of switching the degree of edge enhancement. It is known to determine the processing order of gradation conversion and edge enhancement. (See Patent Document 1).

Further, for example, in order to adjust the edge enhancement degree according to the density without performing the density determination process, the scene of the processing target image is determined based on the shooting mode selection operation or the like, and the density range in which the γ value changes is set. It is known to select a plurality of edge enhancement gradation characteristics correspondingly and perform edge enhancement after edge enhancement gradation conversion processing. (See Patent Document 2).
JP 2002-247410 A JP 2002-281348 A

  According to Patent Documents 1 and 2 described above, after specifying a scene (image) to be processed, enhancement processing corresponding to the type of the scene is applied. However, the enhancement gain is a fixed value for the entire image and for temporal changes of the image.

  For this reason, as shown by a dotted line in FIG. 10A, the post-processing output signal may be larger than the display dynamic range in some or all of the pixels in the screen. That is, the signal value may be saturated. As a result, the image after the edge emphasis process becomes an image without a large contrast difference between white and black, which should be obtained by the process, and becomes an unclear image. Conversely, as indicated by a dotted line in FIG. 10B, the post-processing output signal may use only a part of the display dynamic range in all the pixels in the screen. That is, an unused level may remain. As a result, the image after the edge enhancement processing is an image with a small difference in contrast between white and black compared to an image using the entire display dynamic range. Therefore, the enhancement gain is not optimized, resulting in an image that is not beautiful or not good for the viewer. 10A and 10B correspond to FIG. 3A.

  According to Patent Documents 1 and 2 described above, a scene to be processed is divided into a plurality of patterns (portraits, landscapes, night views, slow syncs, macros, characters, etc.), and an emphasis gain is set for each pattern. Change the value. However, even in this case, the enhancement gain for the divided scenes is a fixed value. For this reason, as described above, the signal in the image after the edge enhancement processing (image after the edge enhancement processing) becomes larger than the display dynamic range or only a part of the display dynamic range is used. .

  As a method for setting the emphasis gain, a technique is also known in which the user manually selects an emphasis gain (value) from several values for an input image. However, in this case, every time the input scene changes, the user must manually reset the enhancement gain, which is troublesome.

  The present invention provides an image signal processing apparatus that can obtain an edge-enhanced image that effectively uses the display dynamic range within a range that does not reach the saturation level by dynamically changing the enhancement gain according to the image. With the goal.

  The present invention also provides an image signal processing method that can obtain an edge-enhanced image that effectively uses the display dynamic range within a range that does not reach the saturation level by dynamically changing the enhancement gain according to the image. The purpose is to do.

  The image signal processing device of the present invention is a local average value calculation unit that calculates a local average value for each pixel of an image to be processed, and a difference between the input value and the local average value for each pixel. A local component value calculation unit that calculates a local component value, and a local gain calculation storage unit that calculates, for each pixel, a local gain that is a value obtained by dividing a difference between a display dynamic range and the local average value by the local component value An enhancement gain calculation unit that uses a minimum value among the local gains for each pixel as an enhancement gain for the image to be processed, and the input value and the local average value for each pixel of the image to be processed And an edge enhancement processing unit that obtains an edge enhanced image of the image to be processed by adding the local average value to a value obtained by multiplying the difference by the enhancement gain.

  Preferably, the image signal processing apparatus of the present invention further includes an edge region determination unit that extracts an edge region from the image to be processed, and the pixels belonging to the edge region extracted by the edge region determination unit are The local component value calculation unit calculates the local component value, and the local gain calculation storage unit calculates the local gain.

  Preferably, the image signal processing apparatus according to the present invention further includes a luminance distribution calculation unit that calculates a luminance distribution for an image to be processed, and the processing based on the luminance distribution calculated by the luminance distribution calculation unit. A noise region determination unit that extracts a noise region from the target image, wherein the enhancement gain calculation unit determines the local gain for pixels belonging to the noise region extracted by the noise region determination unit as the enhancement gain candidate. Exclude from

  Preferably, in the image signal processing device according to the present invention, when the image to be processed further includes bright spot noise, the local gain of the pixel belonging to the region where the bright spot noise exists is determined as the enhancement gain. A local gain determination unit that is excluded from the candidates.

  The image signal processing method of the present invention calculates a local average value for each pixel of an image to be processed, and calculates a local component value that is a difference between the input value and the local average value for each pixel. Calculating a local gain that is a value obtained by dividing a difference between a display dynamic range and the local average value by the local component value for each pixel, and calculating a minimum value in the local gain for each pixel as the processing target. By adding the local average value to the value obtained by multiplying the difference between the input value and the local average value by the enhancement gain for each pixel of the image to be processed, An edge-enhanced image of the image to be processed is obtained.

  According to the image signal processing device and the image signal processing method of the present invention, by using a local component value that is a difference between an input value and a local average value, the enhancement gain is set and the enhancement gain is dynamically changed according to the image. Change to Optimize. Thereby, it is possible to prevent the signal from becoming larger than the display dynamic range in some or all of the pixels in the screen. Further, it is possible to prevent the signal from using only a part of the display dynamic range in all the pixels on the screen. Therefore, an edge-enhanced image that effectively uses the display dynamic range can be obtained within a range that does not reach the saturation level. That is, the entire display dynamic range can be used to optimize the contrast between black and white to obtain an image that is beautiful or good for the viewer. On the other hand, it is not necessary to specify or classify a scene to be processed for setting the enhancement gain, and it is not necessary to change the setting value of the enhancement gain for each scene classification. Therefore, it is not necessary for the user to set the emphasis gain, so that the operation load on the user can be reduced.

  Further, according to the image signal processing device of the present invention, the local component value is calculated and the local gain is calculated for the pixels belonging to the extracted edge region. As a result, the number of pixels to be processed can be reduced, and the processing speed can be increased.

  Further, according to the image signal processing device of the present invention, the local gain for pixels belonging to the noise region extracted based on the luminance distribution is excluded from the enhancement gain candidates. Thereby, it is possible to set the enhancement gain while ignoring the local gain for the region that is significantly deviated from the luminance distribution. As a result, an edge-enhanced image that effectively utilizes the display dynamic range can be obtained by preventing the output signal after edge enhancement from being saturated for any pixel.

  Further, according to the image signal processing device of the present invention, the local gain for the pixels belonging to the region where the bright spot noise exists is excluded from the enhancement gain candidates. Thereby, it is possible to prevent the enhancement gain from being set to a remarkably small value due to the local gain resulting from the bright spot noise. As a result, it is possible to prevent only a part of the display dynamic range from being used.

  FIG. 1 is a diagram showing a configuration of an image signal processing apparatus according to an embodiment of the present invention. 2 and 3 are diagrams for explaining edge enhancement processing by enhancement gain setting processing according to the present invention.

  The image signal processing apparatus includes an edge enhancement processing unit 1 and an enhancement gain processing unit 2. The edge enhancement processing unit 1 includes a local average value calculation unit 11, a first adder 12, a multiplier 13, and a second adder 14. An input signal is input to the edge enhancement processing unit 1. The edge enhancement processing unit 1 outputs an output signal obtained by performing edge enhancement on the input signal. The enhancement gain processing unit 2 includes a local component value calculation unit 21, a local gain calculation storage unit 22, and an enhancement gain calculation unit 23. The edge enhancement processing unit 1 inputs an input signal to the local component value calculation unit 21 of the enhancement gain processing unit 2. The enhancement gain calculation unit 23 of the enhancement gain processing unit 2 inputs the enhancement gain H to the multiplier 13 of the edge enhancement processing unit 1.

  When the image signal processing device is provided in an imaging device such as a camera, for example, the input signal is passed through, for example, a CCD, an AD converter, a signal correction circuit, etc. (all not shown), Input to the edge enhancement processing unit 1. As is well known, the output signal is stored in, for example, an image memory, and then output to a display device (both not shown) via a DA converter and a signal processing circuit. The input signal (hereinafter also referred to as input value) is a signal value of a pixel of the input image (processing target image) 3 in FIG. The output signal (hereinafter also referred to as an output value) is a signal value of a pixel of the output image (image after the enhancement gain process) 4 in FIG.

  The local average value calculation unit 11 calculates the local average value mean (i, j) for each pixel P (i, j). In this example, all the pixels of the processing target image are set as processing target pixels (target pixels) P (i, j), for example, in the order of known scanning. The local average value mean (i, j), which is the output of the local average value calculation unit 11, is input to the first adder 12, the second adder 14, and the local component value calculation unit 21.

  In this example, all pixels constituting the processing target image are set as processing target pixels (target pixels) P (i, j). That is, the local average value mean (i, j) is calculated for the input values (input signals) image_in (i, j) of all the pixels. Thereby, the optimum enhancement gain H can be set (changed) in accordance with the image (scene) to be processed. As a result, for example, an edge-enhanced image in which the signal after edge enhancement on the screen is equal to the display dynamic range can be obtained.

  In the enhancement gain processing unit 2, the local component value calculation unit 21 calculates the input value image_in (i, j) and the local average value mean (i, j) for the target pixel P (i, j) (for each pixel). ) Is calculated as a local component value Δimage (i, j) which is a difference from). The local component value Δimage (i, j) that is the output of the local component value calculation unit 21 is input to the local gain calculation storage unit 22. In addition, although not shown, the local component value calculation unit 21 inputs the local average value mean (i, j) input from the local average value calculation unit 11 to the local gain calculation storage unit 22.

  The local gain calculation storage unit 22 obtains a difference between the display dynamic range and the local average value mean (i, j) for each pixel of interest P (i, j) (for each pixel), and further calculates this difference as a local component value. A local gain Δlocal_gain (i, j), which is a value divided by Δimage (i, j), is calculated and stored. When the calculation of the local gain Δlocal_gain (i, j) for all the pixels of one screen to be processed is completed, the local gain calculation storage unit 22 stores the local gain Δlocal_gain (i for one screen that has been stored. , j) is input to the emphasis gain calculation unit 23.

  The enhancement gain calculation unit 23 sets the minimum value in the local gain Δlocal_gain (i, j) for the target pixel P (i, j) as the enhancement gain H for the one screen (input image 3). That is, the minimum value is obtained from the local gains Δlocal_gain (i, j) for one screen to be processed, and this is set as the enhancement gain H for one screen to be processed. Therefore, this enhancement gain H is applied to all the pixels of one screen that are the processing target.

  For each pixel of the image to be processed, the edge enhancement processing unit 1 locally applies a value obtained by multiplying the difference between the input value image_in (i, j) and the local average value mean (i, j) by the enhancement gain H. By adding the average value mean (i, j), an edge-enhanced image of the image to be processed is obtained. That is, in the edge enhancement processing unit 1, the first adder 12 calculates the difference between the input value image_in (i, j) and the local average value mean (i, j) for the pixel to be processed, and the multiplier 13 multiplies the difference by the enhancement gain H calculated by the enhancement gain calculation unit 23, and the second adder 14 adds the local average value mean (i, j) to the multiplication result obtained by the multiplier 13. As a result, an output signal image_out is obtained as the output of the second adder 14. By repeating the above processing for all the pixels of one screen to be processed, an edge-enhanced image that fully utilizes the display dynamic range can be obtained within a range that does not reach the saturation level.

  Hereinafter, details of the calculation processing of the enhancement gain H and the edge enhancement processing using the enhancement gain H according to the present invention will be specifically described with reference to FIG. 2 and FIG.

  As shown in FIG. 2A, the local average value calculation unit 11 calculates the local average value mean (i, j) for the input value image_in (i, j) of the target pixel P (i, j). To do. The local average value mean (i, j) is calculated by obtaining an average value of pixels corresponding to the size of the filter (spatial filter) F for the target pixel P (i, j). Accordingly, the calculation of the local average value mean (i, j) of the pixel of interest P (i, j) requires input signals (image data) for the size of the filter F. The local average value calculation unit 11 calculates the local average value mean (i, j) at the timing when input of the input signal corresponding to the size of the filter F is completed.

  As shown in FIG. 2B, the filter F includes, for example, three (m = 3) regions in the vertical and horizontal directions with the target pixel P (i, j) as the center. In general, m is an odd number, but may be four (m = 4), that is, an even number as shown in FIG. By such edge enhancement processing by the spatial filter, edge enhancement can be performed for both the horizontal and vertical edges of the input image 2.

  On the other hand, the size of the screen in the input image 3 is, for example, (M + 1) pixels in the horizontal direction of the screen and (N + 1) pixels in the vertical direction of the screen. By scanning the input image 3 in a predetermined direction as is well known, the input value image_in of each pixel is input in a predetermined order. The local average value calculation unit 11 holds the input value image_in as much as necessary for the calculation of the local average value mean (i, j) during the necessary period in units of rows.

  Focusing on the size of the filter F, as shown in FIG. 2B, when m is an odd number, the pixel of interest P (i, j) exists at the center of the filter F. Therefore, the local average value mean (i, j) is calculated by the following equation.

  On the other hand, focusing on the size of the filter F, as shown in FIG. 2C, when m is an even number, the local average value mean (i, j) is calculated as follows. In this case, the position of the target pixel P (i, j) in the filter F is biased in any direction. FIG. 2C illustrates a case where the position of the target pixel P (i, j) is biased to the upper left. Therefore, the local average value mean (i, j) is calculated according to one of the following equations according to the position bias of the target pixel P (i, j) in the filter F.

  When the local average value mean (i, j) is conceptually represented, it is indicated by a thick dotted line in FIG. That is, the local average value mean (i, j) is obtained by distributing (averaging) the change in the input signal image_in to the surrounding predetermined pixels in the edge region of the input signal image_in. Note that FIG. 3A shows a change in the input value (input signal) image_in in one row (or one column) when the input image 3 is scanned in a predetermined direction. In FIG. 3A, the horizontal axis is the pixel position, and the vertical axis is the luminance level of the pixel. FIG. 3A shows an edge where the input signal image_in changes from a small value (before the edge) to a large value (after the edge). A region where the input signal image_in is a small value is called before the edge, and a region where the input signal image_in is a large value is called after the edge. For the opposite edge in FIG. 3A, the local average value mean (i, j) can be obtained in the same manner.

  Next, the local component value calculation unit 21 determines the local component value Δimage (i) which is the difference between the input value image_in (i, j) of the target pixel P (i, j) and the local average value mean (i, j). , j). Further, the local component value calculation unit 21 calculates a difference Δdr (i, j) between the display dynamic range and the local average value mean (i, j). At this time, the sign of the local component value Δimage (i, j) and the sign of the difference Δdr (i, j) are both positive before and after the edge of the image to be processed. The calculation process of the difference Δdr (i, j) may be performed by the local gain calculation storage unit 22.

When the input value image_in (i, j) of the target pixel P (i, j) is larger than the local average value mean (i, j), the local component value Δimage (i, j) is expressed by the following equation Δimage (i, j) ) = image_in (i, j) − mean (i, j)
Is calculated by Further, the difference Δdr (i, j) between the maximum display level dr_max and the local average value mean (i, j) is expressed by the following equation Δdr (i, j) = dr_max−mean (i, j)
Is calculated by Here, the maximum display level dr_max is 256 (gradation) when the display dynamic range is represented by 8-bit gradation.

On the other hand, when the input value image_in (i, j) of the target pixel P (i, j) is smaller than the local average value mean (i, j), the local component value Δimage (i, j) is expressed by the following expression Δimage (i , j) = mean (i, j) − image_in (i, j)
Is calculated by Further, the difference Δdr (i, j) between the minimum display level dr_min of the display dynamic range and the local average value mean (i, j) is expressed by the following equation Δdr (i, j) = mean (i, j) −dr_min
Is calculated by The minimum display level dr_min is 0 (gradation) when the display dynamic range is represented by 8-bit gradation.

  The local component value Δimage (i, j) is conceptually shown in FIG. When the input value image_in (i, j) is smaller than the local average value mean (i, j) (that is, before the edge), the input value image_in (i, j) is calculated from the local average value mean (i, j). Is subtracted to calculate the absolute value of the local component value Δimage (i, j). When the input value image_in (i, j) is larger than the local average value mean (i, j) (that is, after the edge), the inverse operation calculates the absolute value of the local component value Δimage (i, j) Is done.

  On the other hand, before the edge, as can be seen from FIG. 3B, the input signal image_in is emphasized in the direction of the minimum display level dr_min of the display dynamic range. Accordingly, as shown in FIG. 3A, the difference Δdr (i between the display dynamic range and the local average value mean (i, j) is obtained by subtracting the minimum display level dr_min from the local average value mean (i, j). , j) is calculated. On the contrary, after the edge, the input signal image_in is emphasized in the direction of the maximum display level dr_max of the display dynamic range. Therefore, the absolute value of the difference Δdr (i, j) between the display dynamic range and the local average value mean (i, j) is calculated by subtracting the local average value mean (i, j) from the maximum display level dr_max. .

Next, the local gain calculation storage unit 22 determines the local gain Δlocal_gain (i, j) as follows: Δlocal_gain (i, j) = Δdr (i, j) / Δimage (i, j)
Calculated by At this time, the local gain calculation storage unit 22 calculates the local gain Δlocal_gain (i, j) so that the sign is positive before and after the edge in the processing target image. That is, the local gain Δlocal_gain (i, j) is calculated such that the local component value Δimage (i, j) and the difference Δdr (i, j) are both positive signs. In practice, these codes are both positive signs by the local component value calculation unit 21 as described above.

  Next, the enhancement gain calculation unit 23 selects the minimum value among the local gains Δlocal_gain (i, j) calculated for each pixel and sets this as the enhancement gain H.

  Next, based on the above processing, the edge enhancement processing unit 1 calculates the output value image_out (i, j) after the edge enhancement processing of the target pixel P (i, j) as

Calculated by That is, this process is repeated for one screen (scene) to be processed. As a result, as shown by the thick dotted line in FIG. 3B, the post-processing output signal (of the edge enhancement process) is within the display dynamic range, and the entire range from the maximum value dr_max to the minimum value dr_min is used. To do. That is, the edge can be emphasized to the maximum and the unused level of the display dynamic range can be eliminated. Thereby, as a screen after the edge emphasis processing (the output image 4 shown in FIG. 2B), it is possible to obtain an image that is beautiful or good for the viewer.

  FIG. 4 is a block diagram showing another example of the configuration of the image signal processing apparatus of the present invention.

  The image signal processing apparatus in FIG. 1 uses the minimum value among the local gains Δlocal_gain (i, j) calculated for all the pixels of the processing target image (scene) as the enhancement gain H for the processing target image. And However, if the size of the image to be processed is large, the processing time becomes extremely long due to the large amount of processing, and real-time processing may not be possible. In this example, in order to avoid such an increase in processing time, one or a plurality of edges (that is, edge regions) in one screen to be processed are extracted, and the local gain Δlocal_gain (i for pixels belonging to the edge region is extracted. , j) and the minimum value among them is determined as the enhancement gain H for the one screen.

  The image signal processing apparatus of this example further includes an edge area determination unit 24 and a second local average value calculation unit 25 in addition to the configuration shown in FIG. The second local average value calculation unit 25 has the same configuration as the local average value calculation unit 11. The edge region determination unit 24 extracts an edge region from the processing target image. This edge region extraction is performed by, for example, a well-known edge extraction process. The second local average value calculation unit 25 calculates the local average value mean (i, j) for the pixels belonging to the edge region among the pixels of the image to be processed.

  The local component value calculation unit 21 uses the local average value mean (i, j) calculated by the second local average value calculation unit 25 for calculation of the local component value Δimage (i, j). As a result, the pixel belonging to the edge region extracted by the edge region determination unit 24 is used as a pixel to be processed, and the local component value Δimage (i, j) is calculated for these pixels, and the local gain Δlocal_gain (i, j) is calculated. And an enhancement gain is calculated based on this. Accordingly, the processing amount for calculating the local component value Δimage (i, j), the processing amount for calculating the local gain Δlocal_gain (i, j), and the processing amount for calculating the enhancement gain are greatly reduced. The processing speed can be improved.

  FIG. 5 is a block diagram showing still another example of the configuration of the image signal processing apparatus of the present invention. In this example, as in the example of FIG. 4, in order to avoid an increase in processing time, an edge region is extracted, a local gain Δlocal_gain (i, j) is calculated for pixels belonging to the edge region, and the minimum value among them is calculated. Is determined as the enhancement gain H for the image to be processed.

  The image signal processing apparatus of this example further includes an edge region determination unit 24 and a local average value storage unit 26 in addition to the configuration shown in FIG. The edge region determination unit 24 extracts an edge region from the processing target image. The local average value storage unit 26 stores the local average value mean (i, j) calculated by the local average value calculation unit 11 for calculation by the local component value calculation unit 21.

  The local component value calculation unit 21 extracts the local average value mean (i, j) for the pixels belonging to the edge region from the local average value mean (i, j) stored by the local average value storage unit 26, and extracts the extracted value. The local average value mean (i, j) is used to calculate the local component value Δimage (i, j). As a result, the pixel belonging to the edge region extracted by the edge region determination unit 24 is used as a pixel to be processed, and the local component value Δimage (i, j) is calculated for these pixels, and the local gain Δlocal_gain (i, j) is calculated. And an enhancement gain is calculated based on this. Accordingly, the processing amount for calculating the local component value Δimage (i, j), the processing amount for calculating the local gain Δlocal_gain (i, j), and the processing amount for calculating the enhancement gain are greatly reduced. The processing speed can be improved.

  FIG. 6 is a block diagram showing still another example of the configuration of the image signal processing apparatus of the present invention, and FIG. 7 is an explanatory diagram of edge enhancement processing in the image signal processing apparatus of FIG.

  The image to be processed may include bright spot noise N as shown in FIG. The bright spot noise N is a noise with extremely high luminance, and has a luminance that greatly exceeds the maximum display level dr_max of the display dynamic range. As described above, when the minimum value of the local gains Δlocal_gain (i, j) is the enhancement gain H, the local gain Δlocal_gain (i, j) calculated for the bright spot noise N is the target to be processed. Is set as the enhancement gain H of the image (screen). In this case, the enhancement gain H is suppressed to a remarkably small value. In this example, in order to avoid such influence of the bright spot noise N, the luminance distribution of pixels for one screen to be processed is obtained, and the enhancement gain H is determined by ignoring pixels that are significantly out of the distribution. To do.

  The image signal processing apparatus of this example further includes a luminance distribution calculation unit 27 and a noise region determination unit 28 in addition to the configuration shown in FIG. The luminance distribution calculation unit 27 calculates the luminance distribution for the processing target image. The noise area determination unit 28 extracts a noise area from the processing target image based on the luminance distribution calculated by the luminance distribution calculation unit 27.

  For example, the luminance distribution calculation unit 27 examines the luminance of all pixels in the processing target image and creates a histogram thereof. Based on the luminance histogram, the noise region determination unit 28 sets one or a plurality of pixels showing an abnormal distribution as a noise region. As described above, since the luminance of the bright spot noise N is an abnormal value, the distribution position of the bright spot noise N in the histogram is also abnormal. That is, the distribution position of the bright spot noise N is separated from the distribution positions of other pixels and becomes an isolated point. By utilizing this, a region where the bright spot noise N exists, that is, a noise region can be extracted.

  The enhancement gain calculation unit 23 excludes the local gain Δlocal_gain (i, j) for the pixels belonging to the noise region extracted by the noise region determination unit 28 from the enhancement gain H candidates. That is, even if the local gain Δlocal_gain (i, j) for a pixel belonging to the noise region is the minimum value, this is not used as the enhancement gain H. As a result, an edge-enhanced image that fully utilizes the display dynamic range can be obtained within a range in which the output signal after the enhancement gain process does not reach the saturation level for any pixel of the image to be processed.

  FIG. 9 is a block diagram showing still another example of the configuration of the image signal processing apparatus of the present invention.

  As described above with reference to FIG. 7, when the processing target image includes the bright spot noise N, the local gain Δlocal_gain (i, j) calculated for the bright spot noise N is always the minimum value of the local gain. Therefore, the enhancement gain H is suppressed to a remarkably small value. In this example, in order to avoid such influence of the bright spot noise N, the distribution of the local gain Δlocal_gain (i, j) of the pixels for one screen to be processed is obtained, and the local gain Δlocal that is significantly deviated from the distribution is obtained. Ignore _gain (i, j) and determine the enhancement gain H.

  The image signal processing apparatus of this example further includes a local gain determination unit 29 in addition to the configuration shown in FIG. If the image to be processed includes bright spot noise N, the local gain determination unit 29 calculates the local gain Δlocal_gain (i, j) for the pixel belonging to the region where the bright spot noise N exists as the enhancement gain H. Exclude from candidates. That is, even if the local gain Δlocal_gain (i, j) for a pixel belonging to the noise region is the minimum value, this is not used as the enhancement gain H. For this purpose, the local gain determination unit 29 is provided in the enhancement gain calculation unit 23.

  For example, the local gain determination unit 29 examines local gains Δlocal_gain (i, j) of all pixels in the processing target image, creates a histogram thereof, and creates a histogram for the local gain Δlocal_gain (i, j). Based on this, one or a plurality of pixels showing an abnormal distribution is set as a noise region. As described above, since the luminance of the bright spot noise N is an abnormal value, the distribution position of the local gain Δlocal_gain (i, j) in the histogram is also abnormal. That is, the distribution position of the bright spot noise N is separated from the distribution positions of other pixels and becomes an isolated point. Using this, the local gain Δlocal_gain (i, j) in the region where the bright spot noise N exists can be extracted.

  The enhancement gain calculation unit 23 excludes the local gain Δlocal_gain (i, j) extracted by the local gain determination unit 29 from the enhancement gain H candidates. That is, even if the extracted local gain Δlocal_gain (i, j) is the minimum value, it is not set as the enhancement gain H. As a result, an edge-enhanced image that fully utilizes the display dynamic range can be obtained within a range in which the output signal after the enhancement gain process does not reach the saturation level for any pixel of the image to be processed.

  As mentioned above, although this invention was demonstrated according to the embodiment, this invention can be variously deformed within the scope of the gist.

  For example, in the above description, the input image 3 is described as an image from the imaging device 5 such as a camera or a video. However, the image to be processed may be an image of one window in a known multi-window system. . That is, as shown in FIG. 8, in the case of a multi-window system in which a plurality of windows are displayed on the display device 6, the input image 3 is composed of, for example, one window 61 (image data thereof). That is, when one screen is composed of a plurality of windows 61, 62,..., Local gains are calculated for all the pixels constituting each of the plurality of windows 61, 62,. In this case, for the window 62 that is partially hidden, for example, the local gain is calculated only for the region displayed on the display device 6 and the enhancement gain is optimized. Note that the local gain may be calculated for the entire window 62 including the region not displayed on the display device 6 to optimize the enhancement gain. As a result, an edge-enhanced image in which the enhancement gain is optimized can be obtained for each of the plurality of windows 61, 62,. As a result, for each of the plurality of windows 61, 62,..., An optimum edge-enhanced image using the display dynamic range can be obtained within a range that does not reach the saturation level.

  The image signal processing apparatus of the present invention can be provided in various image processing apparatuses. For example, the image signal processing apparatus of the present invention may be provided in an imaging apparatus such as a camera or a video, may be provided in an image processing apparatus such as a copying machine, a facsimile, a scanner, or a printer, and is provided in a computer or the like. May be. Any device that performs edge enhancement using a filter (spatial filter) can be provided in any image processing device.

  From the above, the features of the embodiment of the present invention can be grasped as follows.

(Supplementary Note 1) A local average value calculating unit that calculates a local average value for each pixel of an image to be processed;
A local component value calculation unit that calculates a local component value that is a difference between the input value and the local average value for each pixel;
For each pixel, a local gain calculation storage unit that calculates a local gain that is a value obtained by dividing a difference between a display dynamic range and the local average value by the local component value;
An enhancement gain calculation unit that sets the minimum value in the local gain for each pixel as an enhancement gain for the image to be processed;
For each pixel of the image to be processed, an edge-enhanced image of the image to be processed is obtained by adding the local average value to a value obtained by multiplying the difference between the input value and the local average value by the enhancement gain. And an edge enhancement processing unit for obtaining the image signal processing apparatus.

(Supplementary Note 2) The image signal processing apparatus further includes:
An edge region determination unit that extracts an edge region from an image to be processed;
Note that the local component value calculation unit calculates the local component value for pixels belonging to the edge region extracted by the edge region determination unit, and the local gain calculation storage unit calculates the local gain. The image signal processing apparatus according to 1.

(Supplementary Note 3) The image signal processing apparatus further includes:
A second local average value calculating unit for calculating a local average value of the pixels belonging to the edge region;
The image signal processing apparatus according to appendix 2, wherein the local component value calculation unit uses the local average value calculated by the second local average value calculation unit for the calculation of the local component value.

(Supplementary Note 4) The image signal processing apparatus further includes:
A local average value storage unit for storing the local average value;
The local component value calculation unit extracts a local average value for pixels belonging to the edge region from a local average value stored by the local average value storage unit, and uses the extracted local average value for calculation of the local component value. The image signal processing device according to attachment 2, wherein the image signal processing device is used.

(Supplementary Note 5) The image signal processing apparatus further includes:
A luminance distribution calculation unit for calculating a luminance distribution for the image to be processed;
A noise region determination unit that extracts a noise region from the processing target image based on the luminance distribution calculated by the luminance distribution calculation unit;
The image signal processing apparatus according to appendix 1, wherein the enhancement gain calculation unit excludes the local gain for pixels belonging to the noise region extracted by the noise region determination unit from the enhancement gain candidates. .

(Supplementary Note 6) The image signal processing apparatus further includes:
When the image to be processed includes bright spot noise, the image processing apparatus includes a local gain determination unit that excludes the local gain for pixels belonging to a region where the bright spot noise exists from the enhancement gain candidates. The image signal processing apparatus according to appendix 1.

(Additional remark 7) The said local component value calculation part calculates the said local component value so that the code | symbol may become positive before and behind the edge in the image of a process target. Image signal processing of Additional remark 1 characterized by the above-mentioned. apparatus.

(Additional remark 8) The said local gain calculation preservation | save part calculates the said local gain so that the code | symbol may become positive before and behind the edge in the process target image. The image signal processing apparatus of Additional remark 7 characterized by the above-mentioned. .

(Supplementary note 9) The image signal processing device according to supplementary note 1, wherein the image to be processed includes an image of one window in a multi-window system.

(Supplementary Note 10) For each pixel of the image to be processed, the local average value is calculated,
For each pixel, calculate a local component value that is the difference between the input value and the local average value,
For each pixel, calculate a local gain that is a value obtained by dividing the difference between the display dynamic range and the local average value by the local component value,
The minimum value in the local gain for each pixel is set as an enhancement gain for the image to be processed,
For each pixel of the image to be processed, an edge-enhanced image of the image to be processed is obtained by adding the local average value to a value obtained by multiplying the difference between the input value and the local average value by the enhancement gain. An image signal processing method characterized by comprising:

  As described above, according to the present invention, in the image signal processing apparatus and the image signal processing method, the enhancement gain can be dynamically changed and optimized according to the image. As a result, an edge-enhanced image that effectively uses the display dynamic range can be obtained within a range that does not reach the saturation level, and the operation load on the user for setting the enhancement gain can be eliminated. Therefore, in various image processing apparatuses, edge enhancement processing corresponding to the image can be performed without burden on the viewer, and thereby a beautiful or good-looking image can be obtained for the viewer.

It is a block diagram which shows an example of a structure of the image signal processing apparatus of this invention. It is explanatory drawing of the edge emphasis process in the image signal processing apparatus of FIG. It is explanatory drawing of the edge emphasis process in the image signal processing apparatus of FIG. It is a block diagram which shows another example of a structure of the image signal processing apparatus of this invention. It is a block diagram which shows another example of a structure of the image signal processing apparatus of this invention. It is a block diagram which shows another example of a structure of the image signal processing apparatus of this invention. It is explanatory drawing of the edge emphasis process in the image signal processing apparatus of FIG. It is explanatory drawing of the image signal processing apparatus of this invention. It is a block diagram which shows another example of a structure of the image signal processing apparatus of this invention. It is explanatory drawing of the problem of the conventional edge emphasis process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Edge emphasis processing part 2 Enhancement gain process part 11 Local average value calculation part 12 1st adder 13 Multiplier 14 2nd adder 21 Local component value calculation part 22 Local gain calculation preservation | save part 23 Enhancement gain calculation part

Claims (5)

For each pixel of the image to be processed, a local average value calculation unit that calculates the local average value,
A local component value calculation unit that calculates a local component value that is a difference between the input value and the local average value for each pixel;
For each pixel, a local gain calculation storage unit that calculates a local gain that is a value obtained by dividing a difference between a display dynamic range and the local average value by the local component value;
An enhancement gain calculation unit that sets the minimum value in the local gain for each pixel as an enhancement gain for the image to be processed;
For each pixel of the image to be processed, an edge-enhanced image of the image to be processed is obtained by adding the local average value to a value obtained by multiplying the difference between the input value and the local average value by the enhancement gain. And an edge enhancement processing unit for obtaining the image signal processing apparatus. The image signal processing device further includes:
An edge region determination unit that extracts an edge region from an image to be processed;
The local component value calculation unit calculates the local component value for pixels belonging to the edge region extracted by the edge region determination unit, and the local gain calculation storage unit calculates the local gain. Item 2. The image signal processing device according to Item 1. The image signal processing device further includes:
A luminance distribution calculation unit for calculating a luminance distribution for the image to be processed;
A noise region determination unit that extracts a noise region from the processing target image based on the luminance distribution calculated by the luminance distribution calculation unit;
The image signal processing according to claim 1, wherein the enhancement gain calculation unit excludes the local gain for pixels belonging to the noise region extracted by the noise region determination unit from the enhancement gain candidates. apparatus. The image signal processing device further includes:
When the image to be processed includes bright spot noise, the image processing apparatus includes a local gain determination unit that excludes the local gain for pixels belonging to a region where the bright spot noise exists from the enhancement gain candidates. The image signal processing apparatus according to claim 1. For each pixel of the image to be processed, calculate its local average value,
For each pixel, calculate a local component value that is the difference between the input value and the local average value,
For each pixel, calculate a local gain that is a value obtained by dividing the difference between the display dynamic range and the local average value by the local component value,
The minimum value in the local gain for each pixel is set as an enhancement gain for the image to be processed,
For each pixel of the image to be processed, an edge-enhanced image of the image to be processed is obtained by adding the local average value to a value obtained by multiplying the difference between the input value and the local average value by the enhancement gain. An image signal processing method characterized by comprising:

Images