JP2015169984A - Image processor, image processing method, image processing program and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the "distinctness" and the "feeling of unevenness" of an image while suppressing a change in the brightness and the color tone of the image.SOLUTION: An image processor includes a generation part for respectively generating a plurality of resolution images having different resolutions from an input image, an adjustment part for adjusting respective contrasts of the plurality of resolution images by using predetermined adjustment amounts in accordance with respective values of signal values showing pixels of the resolution images to respectively adjust the signal values, a reconfiguration part for reconfiguring an output image corresponding to the input image by respectively using the plurality of resolution images whose contrasts are adjusted by the adjustment part, and an average value storage part for performing processing for matching the average value of contrasts of the output image to the average value of contrasts of the input image.

Description

  The present invention relates to an image processing apparatus, an image processing method, an image processing program, and a recording medium.

  Image data taken by various imaging devices such as digital cameras and scanners, and image data created by image editing software, etc., are subjected to image processing to create a “clearness” and “concaveness”. An image processing apparatus that improves the image quality is known. In other words, there is a problem that the dynamic range that can be covered by the imaging device and the display device is too narrow to represent the real world.

  The following describes what an image lacks in a sense of clarity and unevenness. When a photographer captures a landscape using an imaging device such as a digital camera and reproduces it using a display device such as a display, the captured image is compared with the image in the memory actually viewed at the shooting site. I sometimes feel like a hazy / haze effect. The inventor expresses this as “lack of clarity”.

  Similarly, the unevenness (for example, the state of the trees in the landscape image and the rocky surface of the landscape image) expressed in the image in memory seems to be flat in the photographed image. I sometimes remember the impression. The inventor expresses this as “lack of unevenness”.

  Many kinds of techniques have been proposed for a long time as a technique that contributes to the improvement of the lack of clarity and unevenness. Typical examples include contrast enhancement using a tone curve and an unsharp mask. However, with the recent technological progress, the effect obtained by such a method used for a long time is inevitably inadequate.

  Specifically, when contrast enhancement using a tone curve is applied to an image that already uses the dynamic range of the display device, contrast cannot be enhanced without causing side effects. The side effect here refers to, for example, the inability to avoid the occurrence of a portion where the contrast is reduced without being emphasized. In addition, the unsharp mask has a problem that an edge portion having a large density gradient is excessively emphasized, but a feeling of unevenness corresponding to a small density gradient is difficult to be emphasized.

  In recent years, image processing using a multi-resolution image is known as a technique that contributes to improvement of clarity and unevenness. As a method of generating a multi-resolution image, for example, there is a method of repeating a widely known discrete wavelet transform.

  In Patent Document 1, an image signal representing an original image is converted into a luminance signal and a color difference signal, and the luminance signal and the color difference signal are converted into a multi-resolution of levels 1 to N, respectively. A resolution signal is generated, a high-frequency signal of level 1 of the color-difference multi-resolution signal is suppressed, and then inverse multi-resolution conversion is performed to obtain a processed color-difference signal. The conditions differ for each level of the high-frequency signal of each level of the luminance multi-resolution signal An image processing method for obtaining a processed image signal by synthesizing the processed luminance signal and the processed chrominance signal by performing inverse multi-resolution conversion after performing coring processing using the image to generate a processed luminance signal is disclosed. Yes.

  Patent Document 2 shows an image obtained by generating a basic structural component image indicating an image obtained by smoothing other than the edges of the input image while preserving the edges of the input image, and subtracting the basic structural component image from the input image. Image processing that generates a detailed component image, decomposes the generated detailed component image into a plurality of frequency band images, performs noise removal, and performs multi-resolution inverse transform that combines the noise band-removed frequency band images An apparatus is disclosed.

  However, the conventional technique has a problem that the brightness and color of the image change when image processing is performed to improve the “clearness” and “unevenness”. For example, when conventional image processing is performed to improve “clearness” and “unevenness”, the average value of the pixel values of the input image differs from the average value of the pixel values of the output image. When the input image changes, the deviation amount of the average value also changes. According to the inventor's study, it is considered that an overview (overall impression) of brightness and color of an image is determined by an average value of the image data (pixel value).

  The present invention has been made in view of the above, and an image processing apparatus and image that can improve the “clearness” and “concaveness” of an image while suppressing changes in brightness and color of the image It is an object to provide a processing method, an image processing program, and a recording medium.

  In order to solve the above-described problems and achieve the object, the present invention provides a generation unit that generates a plurality of resolution images having different resolutions from an input image, and a signal value that indicates a pixel of the resolution image. By adjusting each of the signal values using a predetermined adjustment amount, an adjustment unit that adjusts the contrast of each of the plurality of resolution images, and each of the plurality of resolution images that the adjustment unit has adjusted the contrast. And a reconstructing unit that reconstructs an output image corresponding to the input image, and an average value storage unit that performs processing for matching the average value of the contrast of the output image with the average value of the contrast of the input image. It is characterized by that.

  According to the present invention, it is possible to improve the “clearness” and “unevenness” of an image while suppressing changes in brightness and color of the image.

FIG. 1 is a block diagram showing the configuration of the image processing apparatus according to the first embodiment. FIG. 2 is a graph showing a first embodiment of an adjustment amount predetermined according to each signal value. FIG. 3 is a block diagram illustrating a configuration of the image processing apparatus according to the second embodiment. FIG. 4 is a graph showing a second embodiment of the contrast adjustment amount performed by the multi-resolution image operation unit.

  Hereinafter, an embodiment of an image processing apparatus will be described in detail with reference to the accompanying drawings.

(First embodiment)
FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus 10 according to the first embodiment. First, an outline of the image processing apparatus 10 will be described. As shown in FIG. 1, the image processing apparatus 10 includes a multi-resolution image generation unit (generation unit) 20, an average value calculation unit 30, a multi-resolution image operation unit (adjustment unit) 40, an average value storage unit 50, and a reconstruction unit. 60. In addition, each part which comprises the image processing apparatus 10 may be comprised by hardware, and may be comprised by the software performed by CPU which is not shown in figure.

  In the image processing apparatus 10, when image data in a general format is input, the generation unit 20 generates a multi-resolution image from the input image data. The average value calculation unit 30 calculates the average value of the contrast of each resolution image of the multi-resolution image generated by the generation unit 20, and calculates the calculated average value (average value of each resolution image before operation) as an average value storage unit. 50 is output.

  The adjustment unit 40 operates (adjusts) a signal value indicating each pixel of each resolution image of the multiple resolution image generated by the generation unit 20. Specifically, the adjustment unit 40 adjusts the signal value using a predetermined adjustment amount (see FIG. 2) according to the value of each signal value indicating the pixel of the resolution image, thereby Adjust the contrast of each resolution image.

  The average value storage unit 50 stores the average value of the signal values of each resolution image by using the average value calculated by the average value calculation unit 30 for each resolution image whose signal value has been adjusted by the adjustment unit 40. (The signal value is changed). That is, the average value storage unit 50 performs processing for matching the average value of the contrast of the output image with the average value of the contrast of the input image.

  The reconstruction unit 60 reconstructs the output image corresponding to the input image so that the image data has the same format as the input image, using each resolution image processed by the average value storage unit 50.

  In the first embodiment of the image processing apparatus 10, the input image data is, for example, in the TIF format, has three color components of RGB, and has 16-bit data for each color per pixel. The file format of the input image data is not limited to the TIF format, and may be another file format such as JPEG or PNG. Also, the color components are not limited to the RGB color space, and may be a color space other than RGB. Also, the data amount per pixel may be a data amount other than 16 bits.

  Further, the image processing apparatus 10 may adjust only the brightness component of each resolution image, for example, so as to improve the “clearness” and “unevenness” of the input image. As a specific example, the image processing apparatus 10 converts an input image having RGB color components into a YCbCr color space by a conversion unit (not shown), and performs image processing described later on only the Y component. In addition, the image processing apparatus 10 converts an output image that has been subjected to image processing, which will be described later, from a YCbCr color space to an RGB color component by a conversion unit (not shown) and outputs the result.

  However, the image processing apparatus 10 is not limited to the conversion to the YCbCr color space, but is a La * b * color space or HSV color that is a color space having other brightness components (or brightness components). You may convert into space etc. The reason why image processing is performed after the image processing apparatus 10 converts to a color space having a brightness component is that the image processing is performed on only one component compared to performing image processing on three components (RGB three components). This is because processing reduces the amount of calculation. Therefore, the image processing apparatus 10 may be configured to perform image processing to be described later for all three components without changing the color space.

  Next, each part which comprises the image processing apparatus 10 is further explained in full detail.

  The generation unit 20 generates resolution images having a plurality of different resolutions from the input image, for example, using a Laplacian pyramid, for the Y component of the YCbCr color space described above.

  The gradient of the image data greatly affects the improvement of “clearness” and “unevenness”. The signal value of the multi-resolution image constituting the Laplacian pyramid corresponds to a characteristic value closely related to the gradient of the input image. The Laplacian pyramid takes each multi-resolution image by taking the difference between the image data that has undergone a series of processing (smoothing, downsizing, upsizing, smoothing) and the image before processing (Laplacian). It is formed by calculating the component image) and repeating the process. Also, using the Laplacian pyramid as a multi-resolution image constitutes image processing by basic processing, and reduces the amount of calculation compared to other multi-resolution images (for example, discrete wavelet transform).

  The generation unit 20 first smoothes the Y component image (assuming that the image size at this time is the same as the input image and the image data of (M, N) size) with a 5 × 5 Gaussian filter. Process. Next, the generation unit 20 creates image data of ½ size in length and width composed of pixel data of even rows and columns from the smoothed image. Hereinafter, this image is referred to as a scale-down image. Also, the top row and the top column are referred to as the 0th row and the 0th column, respectively.

  The image size of the first scale-down image is (M / 2, N / 2). The generation unit 20 repeatedly applies image generation (scale down operation) of 1/2 size in length and width composed of pixel data (signal values) of even columns and rows after the smoothing process to the input image. In other words, the generation unit 20 generates a scale-down image that is ½ size both vertically and horizontally each time a scale-down operation is performed.

  On the other hand, the generation unit 20 creates an image having the same (M, N) size as the original input image size by again doubling the size of the scale-down image. At this time, the generation unit 20 assigns the same signal value as that of the even row / column to the odd row / column for the time being. That is, four pixels having the same pixel data are generated. Thereafter, the generation unit 20 creates image data that has been smoothed by a 5 × 5 Gaussian filter. Hereinafter, this image is referred to as a scale-up image.

  Next, the generation unit 20 generates an image obtained by calculating the difference between the image before the scale-down and the image after the scale-down for each pixel data (signal value). This image will be referred to as a Laplacian component image below.

  Through the above operation, the generation unit 20 generates a Laplacian component image having an initial resolution (level 0 resolution). The Laplacian component of level 0 resolution is (M, N) size image data.

  The generation unit 20 calculates the difference between the scale-down image and the scale-up image after performing the scale-down again, thereby obtaining a Laplacian component at the next resolution (this is a level 1 resolution Laplacian component image and Calculated). The Laplacian component of level 1 resolution is (M / 2, N / 2) size image data.

  Thereafter, the generation unit 20 repeats the scale-down and the calculation of the Laplacian component until the (2, 2) size Laplacian component is calculated. A series of Laplacian component images from a level 0 resolution Laplacian component image ((M, N) size) to a level n resolution Laplacian component image ((2 × 2) size) is a resolution image of each level of the multi-resolution image. Corresponding to The level n is a different value depending on the image size of the input image.

  The generation unit 20 has been described with respect to an example in which a 5 × 5 Gaussian filter is used as a smoothing filter in scale-down and scale-up, but is not limited thereto. The generation unit 20 may perform scale down and scale up by so-called interpolation processing instead of the smoothing filter. Specific examples of the interpolation process include a bilinear method and a bicubic method.

  In addition, the generation unit 20 uses, for example, a so-called Laplacian pyramid for the multi-resolution image, but the multi-resolution image may be generated by other methods without being limited thereto. The image processing apparatus 10 focuses on a difference (gradient) between adjacent pixels of the input image and intends to operate the difference. That is, the generation unit 20 may generate any multi-resolution image as long as the difference (gradient) between adjacent pixels is reflected in each multi-resolution image. For example, although it is not a very general technique, two images with directionality for each resolution image (the first image reflects a difference between adjacent pixels in the x direction, the second image in the y direction) A multi-resolution image may be generated using a difference between adjacent pixels).

  As described above, the average value calculation unit 30 calculates an average value of contrast for each resolution image for each of the plurality of resolution images generated by the generation unit 20, and outputs the average value to the average value storage unit 50.

  The adjustment unit 40 performs an operation of adjusting (converting) the signal value according to the following expression 1 with respect to the multi-resolution image (a series of Laplacian component images).

  FIG. 2 is a graph showing a first embodiment of an adjustment amount (relation before and after adjustment of a signal value according to the above equation 1) determined in advance according to each signal value. The adjustment unit 40 adjusts the contrast of each resolution image using the adjustment amount shown in FIG.

  In FIG. 2, the solid bold line represents the relationship of Equation 1 above, and the dotted line represents the case where the slope is 1 (reference). A dotted line with an inclination of 1 corresponds to a case where the adjustment unit 40 has made no adjustment to each signal value.

  As is clear from FIG. 2, the adjustment unit 40 gives a relatively large change to a signal value having a small absolute value (the absolute value of the Laplacian component is small), and the absolute value of the signal value is adjusted after the adjustment. Try to increase.

  On the other hand, the adjustment unit 40 gives a relatively small change to a signal value having a large absolute value (a large Laplacian component), and increases or decreases the absolute value of the signal value after adjustment. Yes.

  As described above, the adjustment unit 40 adjusts the signal values using adjustment amounts that are determined in advance according to the values of the signal values indicating the pixels of the resolution image, thereby adjusting the contrast of each of the plurality of resolution images. It is adjusted. That is, the multi-resolution image operation unit (adjustment unit) 40 changes the ratio of the signal values before and after the operation according to the magnitude of the signal value (also referred to as the data value or coefficient value of each pixel). Perform the operation.

  Further, the adjustment of the signal value performed by the adjustment unit 40 is not limited to the above-described example. For example, the adjustment unit 40 may adjust the contrast of the multi-resolution image by using a lookup table (LUT) or the like instead of the mathematical formula shown in the above formula 1.

  Note that the dynamic range (Dr: gradation range) of the input image is 1.0 in the case of the Y component of the YCbCr color space. That is, Dr of the input image takes a value of 0.0 to 1.0. Moreover, the range (V) of the signal value adjusted to increase the absolute value of the signal value (L) of each resolution image shown in FIG.

  That is, when the ratio (L / Dr) of the signal value to the gradation range of the input image is at least less than a predetermined threshold (for example, 0.1), the adjustment unit 40 sets the signal value. adjust.

  The reason why the “clearness” and “unevenness” are improved by adjusting the signal value with respect to the range defined by Equation 2 is not necessarily clear, but is considered as follows. . As described above, the dynamic range of an imaging device such as a camera and a display device such as a display is very narrow compared to the actual dynamic range of a subject to be photographed. For this reason, dynamic range compression is performed in imaging by an image sensor and display on a display. At this time, edge portions and gradients having a relatively small density difference are also compressed more strongly than the visually desired compression amount as the dynamic range is compressed. Therefore, it becomes difficult to perceive edge portions and gradients having a relatively small density difference. As a result, it is considered that “clearness” and “concaveness” are deteriorated.

  Therefore, when the adjustment unit 40 adjusts the signal value within the range defined by the above equation 2, image processing is performed in which the edge and the gradient are relatively increased in a region having a relatively small density difference. Therefore, the edge portion and the gradient having a relatively small density difference that deteriorates with compression of the dynamic range can be increased again, so that “clearness” and “unevenness” are improved.

  The average value storage unit 50 applies the signal value of each resolution image output by the adjustment unit 40 so that the average value of the resolution image matches for each resolution level before and after the adjustment unit 40 adjusts the signal value. The process using the following formula 3 is performed.

  The reconstructing unit 60 repeats the operation opposite to that of the generating unit 20, thereby using each of the plurality of resolution images obtained by adjusting the contrast by the adjusting unit 40 and processing the average value by the average value storing unit 50. The output image corresponding to is reconstructed. That is, the reconstructing unit 60 repeats processing at each resolution level by starting computation from the resolution level on the low resolution side and adding the signal values of each resolution level processed by the average value storage unit 50 after scaling up. Reconstruct the component image.

  Further, the reconstruction unit 60 performs the calculation in the same manner as the scale-up in the generation unit 20 also in the scale-up. That is, the reconstruction unit 60 assigns the same signal value as that of the even-numbered rows / columns for the odd-numbered rows / columns to the image having a size doubled vertically and horizontally. Thereafter, the reconstruction unit 60 generates a scale-up image by creating image data that has been smoothed by a 5 × 5 Gaussian filter.

  The image processing apparatus 10 adds one row (column) so that the image size always becomes an even number when the image size becomes an odd number depending on the input image when repeating the scale down. Then scale down. Then, the image processing apparatus 10 performs addition by assigning the same signal value as that of the maximum row (column) of the image to one row (one column) to be added. Further, the image processing apparatus 10 stores in advance at which resolution level the added row (column) is added in this way, and removes the added row (column) when reconstructing. To return to a higher level.

  The image processing apparatus 10 is not limited to always maintaining the image size at an even number. For example, the image processing apparatus 10 may change the image size by complement processing instead of adding one row (column) as described above. The image processing apparatus 10 may generate a multi-resolution image after expanding the image size of the input image to a factorial of 2 in advance.

  As described above, the image processing apparatus 10 performs conversion from the YCbCr color space to the RGB color space using the new Y component data generated by the series of processes, and uses the image data having the RGB color components as an output image. Output. Here, the image processing apparatus 10 uses the CbCr component without any processing after generation from the input image.

  Accordingly, even if the image processing apparatus 10 adjusts the contrast of each of the plurality of resolution images, the average value storage unit 50 causes the average value of the pixel values of the output image to change from the average value of the pixel values of the input image. Therefore, it is possible to improve the “clearness” and “unevenness” of the image while suppressing changes in brightness and color of the image.

(Second Embodiment)
FIG. 3 is a block diagram showing the configuration of the image processing apparatus 10a according to the second embodiment. First, an outline of the image processing apparatus 10a will be described. As shown in FIG. 3, the image processing apparatus 10a includes an average value calculation unit 30a, a multi-resolution image generation unit (generation unit) 20a, a multi-resolution image operation unit (adjustment unit) 40a, a reconstruction unit 60a, and an average value storage unit. 50a. In addition, each part which comprises the image processing apparatus 10a may be comprised with hardware, and may be comprised with the software performed by CPU which is not shown in figure.

  The image processing apparatus 10a according to the second embodiment is configured such that the average value of the contrast before and after the adjustment of the signal value is stored at each resolution level in the image processing apparatus 10 according to the first embodiment. On the other hand, the image is processed so that the average value is stored for the image after a series of image processing (multi-resolution image generation, signal value adjustment, reconstruction). Similarly to the image processing apparatus 10, the image processing apparatus 10 a adjusts contrast only for the Y component in the YCbCr color space.

  The image processing apparatus 10a converts an input image having RGB color components into a YCbCr color space by a conversion unit (not shown), and performs image processing to be described later on only the Y component. Further, the image processing apparatus 10a converts an output image that has been subjected to image processing described later from a YCbCr color space to an RGB color component by a conversion unit (not shown) and outputs the converted image.

  The average value calculation unit 30a calculates an average value for the input image (Y component image), and outputs the calculated average value to the average value storage unit 50a.

  Similar to the generation unit 20, the generation unit 20a generates a multi-resolution image. Similar to the adjustment unit 40, the adjustment unit 40a adjusts the contrast. Similar to the reconstruction unit 60, the reconstruction unit 60a repeats addition of images at each resolution level after scale-up and signal value adjustment to generate a Y component image after contrast adjustment.

  The average value storage unit 50a includes an average value of Y component images reconstructed by the reconstruction unit 60a after adjusting the contrast by the adjustment unit 40a, and an average of Y component images calculated by the average value calculation unit 30a for the input image. Using the value, the average value of the Y component data is stored by the following equation 4.

  In the image processing apparatus 10a according to the second embodiment, the average value storage unit 50a stores the average value of the Y component data after the reconstruction unit 60 reconstructs an image from the multi-resolution image. Is once for one piece of image data (reconstructed image). Therefore, the image processing apparatus 10a according to the second embodiment suppresses a change in the brightness and color of the image with a smaller calculation amount than the image processing apparatus 10 according to the first embodiment, and the “clearness” of the image and It is possible to improve the “unevenness”.

  On the other hand, the image processing apparatus 10 according to the first embodiment stores the average value of the signal values of the respective resolution images and reconstructs the output image. Therefore, the image processing apparatus 10 according to the first embodiment performs image reconstruction more than the image processing apparatus 10a according to the second embodiment. Partial (local) average value changes can be mitigated. That is, the image processing apparatus 10 can better preserve the appearance (overall impression) of the input image and improve the “clearness” and “concaveness”.

  Next, a second embodiment of contrast adjustment performed by the adjustment unit 40 (or adjustment unit 40a) will be described. FIG. 4 is a graph showing a second embodiment of the contrast adjustment amount performed by the adjustment unit 40 (or the adjustment unit 40a). 4B is an enlarged graph showing the central portion (region where the absolute value of the Laplacian component is small) of the graph shown in FIG.

  The second embodiment of the adjustment amount shown in FIG. 4 differs from the first embodiment of the adjustment amount shown in FIG. 2 in the adjustment amount of the signal value at each resolution level. Specifically, in the second embodiment, the amount of adjustment is suppressed more than in the first embodiment in a low gradient region (data location where the absolute value of the Laplacian component is a small value).

  In FIGS. 4A and 4B, the thick solid line indicates the relationship between the Laplacian components before and after adjustment in the second embodiment. A thin solid line indicates a relationship between Laplacian components before and after adjustment in the first embodiment. A dotted line shows the relationship of the inclination 1 as a reference.

  In particular, as can be seen from FIG. 4B, there is a difference in the area where the absolute value of the Laplacian component is less than 0.04 between the first example and the second example, and the absolute value of the Laplacian component is 0.04. The above areas are the same. That is, in the second embodiment, the adjustment amount of the signal value is suppressed as compared with the adjustment amount of the first embodiment in a region where the absolute value of the Laplacian component is less than 0.04.

  Hereinafter, effects obtained in the second embodiment will be described. The noise component included in the image is an object that does not want to be enhanced, unlike the signal component of the image (the light and shade that actually existed in the imaged object). On the other hand, it is very difficult to avoid that noise generated for various reasons is included (superimposed) in the captured image. According to the inventor's investigation, it has been found that the characteristics of such a noise component are often not a large gradient but a very small gradient. Considering such a small gradient from the viewpoint of the Laplacian component, it corresponds to a very small Laplacian component. On the other hand, it has been clarified by examination that the signal component of the image corresponds to a slightly larger Laplacian component than the noise component. The second embodiment reflects such examination results. That is, the adjustment amount of a small Laplacian component that tends to enhance the noise component is suppressed to prevent the noise from being emphasized. As described above, when the second embodiment of the adjustment amount is used, an adjustment that brings the signal component of the image closer to the light and shade existing in the actual shooting target while suppressing noise enhancement is performed. It becomes possible to improve the “unevenness”.

  Next, a third embodiment of contrast adjustment performed by the adjustment unit 40 (or adjustment unit 40a) will be described. In the third embodiment, the signal value adjustment performed in the second embodiment is not applied to all resolution levels, but is applied to both the high resolution side and the low resolution side. The adjustment of the signal value applied to the medium resolution side is the same as in the first embodiment. In other words, on the high resolution side and the low resolution side, the adjustment amount is suppressed in the low gradient region (data location where the absolute value of the Laplacian component is small), and for the medium resolution image, The amount of adjustment is not suppressed.

  The distinction between high resolution, medium resolution, and low resolution in the third embodiment is as follows. In the third embodiment, the number of pixels of the input image is, for example, 4928 × 3280 pix. In the case of this image size, for example, the resolution level generated by the generation unit 20 is from level 0 to level 13. For example, the resolution level on the high resolution side is set to resolution level 0 to resolution level 2. And the process which suppressed the adjustment amount of the low gradient area | region shown in FIG. 4 is applied to the resolution image of the resolution level 0-resolution level 2. FIG.

  Next, the processing that does not suppress the adjustment amount of the low gradient region shown in FIG. 2 is applied with the resolution level 3 to the resolution level 7 as the medium resolution level. Further, the processing from the resolution level 8 to the resolution level 13 is performed by reducing the adjustment amount of the low gradient region shown in FIG. 4 again as the low resolution level.

  The effects obtained in the third embodiment will be described below. As described above, the noise component included in the image is often caused by a very small gradient rather than a large gradient. According to the inventor's study, it has been found that the noise component included in the image is further included in the high-resolution image. On the other hand, it has also been clarified that the medium-resolution image includes the light and shade actually present in the object to be imaged, which is a signal component of the image, even with a very small gradient. At low resolution, there is a problem that the density step (pseudo contour) becomes conspicuous because the amount of adjustment of the low gradient increases rather than noise. Also for this problem, the occurrence of the problem can be solved by suppressing the adjustment amount of the Laplacian component corresponding to a very small gradient.

  In other words, when the third embodiment is used, it suppresses noise enhancement and the occurrence of gradation steps (pseudo contours), while suppressing the signal component of the image (the light and shade that actually existed in the shooting target) more than necessary. The signal component of the image can be adjusted without doing so. As a result, the “clearness” and “unevenness” of the image can be improved.

  An image processing program executed by the image processing apparatus 10 (or the image processing apparatus 10a) of the present embodiment is a file in an installable format or an executable format, and is a CD-ROM, a flexible disk (FD), a CD-R, It is recorded on a computer-readable recording medium such as a DVD (Digital Versatile Disk) and provided as a computer program product.

  Further, the image processing program executed by the image processing apparatus 10 (or the image processing apparatus 10a) of the present embodiment is stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. You may comprise.

  In addition, the image processing program of the present embodiment may be provided by being incorporated in advance in a ROM or the like.

  As described above, according to the image processing device 10 or the image processing device 10a, even when the dynamic range is sufficiently used, all density regions are generated without generating a density region in which the contrast is lowered. It is possible to enhance the contrast. The adjustment unit 40 of the image processing apparatus 10 according to the first embodiment is not limited to performing contrast adjustment using a predetermined adjustment amount illustrated in FIG. Further, the adjustment unit 40a of the image processing apparatus 10a according to the second embodiment is not limited to performing contrast adjustment using a predetermined adjustment amount illustrated in FIG. That is, the image processing apparatus 10 and the image processing apparatus 10a may adjust the contrast using the adjustment amounts shown in FIG. 2 or FIG. 4, respectively, or may be other different predetermined adjustments not shown. The contrast may be adjusted using the amount.

10, 10a Image processing device 20, 20a Multi-resolution image generation unit (generation unit)
30, 30a Average value calculation unit 40, 40a Multi-resolution image operation unit (adjustment unit)
50, 50a Average value storage unit 60, 60a Reconfiguration unit

JP 2003-134352 A Japanese Patent No. 5348145

Claims (8)

A generating unit that generates a plurality of resolution images of different resolutions from the input image;
An adjustment unit that adjusts the contrast of each of the plurality of resolution images by adjusting the signal value using a predetermined adjustment amount according to the value of each of the signal values indicating the pixels of the resolution image;
A reconstructing unit that reconstructs an output image corresponding to the input image using each of the plurality of resolution images whose contrast has been adjusted by the adjusting unit;
An image processing apparatus, comprising: an average value storage unit configured to perform processing for matching an average value of contrast of the output image with an average value of contrast of the input image. For each of the plurality of resolution images generated by the generation unit, an average value calculation unit that calculates an average value of contrast for each resolution image,
The average value storage unit
The average value calculation unit performs a process of matching the average value of the contrast of the output image with the average value of the contrast of the input image, using each of the average values of the contrast calculated for each resolution image. The image processing apparatus according to claim 1. An average value calculating unit for calculating an average value of the contrast of the input image,
The average value storage unit
Using the average contrast value calculated by the average value calculation unit and the output image reconstructed by the reconstruction unit, the average contrast value of the output image is matched with the average contrast value of the input image. The image processing apparatus according to claim 1, wherein processing is performed. The generator is
The image processing apparatus according to any one of claims 1 to 3, wherein a plurality of the resolution images are respectively generated by a Laplacian pyramid. The adjustment unit is
The signal value is adjusted when the ratio of the value of the signal value to the gradation range of the input image is at least less than a predetermined threshold value. The image processing apparatus according to item. Generating a plurality of different resolution images from the input image, and
Adjusting the contrast of each of the plurality of resolution images by adjusting each of the signal values using a predetermined adjustment amount according to the value of each of the signal values indicating the pixels of the resolution image;
Reconstructing an output image corresponding to the input image using each of the plurality of resolution images adjusted in contrast;
A process of adjusting an average contrast value of the output image to an average contrast value of the input image. Generating a plurality of resolution images of different resolutions from the input image,
Adjusting the contrast of each of the plurality of resolution images by adjusting each of the signal values using a predetermined adjustment amount according to the value of each of the signal values indicating the pixels of the resolution image;
Reconstructing an output image corresponding to the input image using each of the plurality of resolution images adjusted in contrast;
An image processing program for causing a computer to execute a process of adjusting an average contrast value of the output image to an average contrast value of the input image. Generating a plurality of resolution images of different resolutions from the input image,
Adjusting the contrast of each of the plurality of resolution images by adjusting each of the signal values using a predetermined adjustment amount according to the value of each of the signal values indicating the pixels of the resolution image;
Reconstructing an output image corresponding to the input image using each of the plurality of resolution images adjusted in contrast;
A computer-readable recording medium having recorded thereon an image processing program for causing a computer to execute a process of adjusting an average contrast value of the output image to an average contrast value of the input image.

Images