JP4760433B2 - Image processing apparatus, image processing method, and program - Google Patents

Description

  The present invention relates to an image processing apparatus, an image processing method, and a program, and more particularly, to an image processing apparatus, an image processing method, and a program that are suitable for use when correcting contrast based on a histogram of an image luminance signal.

  As a method for correcting the contrast of an image, a histogram method using a histogram of luminance signals is known (see, for example, Patent Document 1).

  FIG. 1 shows an example of the configuration of a conventional contrast correction apparatus that improves the contrast of an image by applying a histogram method.

  The contrast correction apparatus 10 receives an image luminance signal Yi and outputs a corrected (improved) luminance signal Yo. A histogram data acquisition unit 11, a cumulative frequency conversion unit 12, and a normalized gain calculation unit 13. , LUT luminance conversion unit 14, and timing control unit 15.

  In the contrast correction apparatus 10, the luminance signal Yi input from the previous stage is input to the histogram data acquisition unit 11 and the LUT luminance conversion unit 14.

  The histogram data acquisition unit 11 counts the histogram data of the luminance signal Yi for one input image (one field or one frame), that is, the frequency of each luminance (appearance frequency of pixels) to convert the cumulative frequency. To the unit 12. The cumulative frequency conversion unit 12 converts the histogram data input from the histogram data acquisition unit 11 into a cumulative frequency for each luminance, that is, an addition value from a frequency corresponding to the minimum luminance value to a frequency corresponding to the luminance, and normalizes it. Output to the gain calculator 13.

  The normalization gain calculation unit 13 performs normalization by dividing the cumulative frequency corresponding to each luminance input from the cumulative frequency conversion unit 12 by the maximum luminance value (for example, 255 if the luminance is 8 bits). The normalized gain curve obtained as a result is output to the LUT luminance conversion unit 14.

  The LUT luminance conversion unit 14 holds the normalized gain curve input from the normalized gain calculation unit 13 as a look-up table (hereinafter referred to as LUT), and the luminance signal Yi input from the previous stage of the contrast correction device 10. , The luminance signal Yo is converted using the held LUT and output to the subsequent stage.

  The timing control unit 15 controls the operation timing of the histogram data acquisition unit 11 to the LUT luminance conversion unit 14.

  According to the contrast correction apparatus 10, the histogram data acquisition unit 11 has, for example, a pixel of an input image of 720 (horizontal) × 480 (vertical), its luminance is 8 bits, and is as shown in FIG. When the histogram data is acquired, the cumulative frequency is converted into the cumulative frequency as shown in FIG. 3 by the cumulative frequency converter 12, and further, the normalized gain calculation unit 13 performs the smooth right upward normalization as shown in FIG. A gain curve is obtained. This normalized gain curve is held in the LUT luminance conversion unit 14 and is referred to when the input luminance signal Yi is converted into the output luminance signal Yo.

  For example, the luminance signal Yi corresponding to the histogram data shown in FIG. 2 is converted into the luminance signal Yo of the histogram shown in FIG. As shown in FIG. 5, since the output luminance signal Yo has a histogram distributed over the entire dynamic range, the contrast of the image is improved.

JP 2002-140701 A

  However, the histogram of the luminance signal of an image that is actually intended to improve contrast is rarely a smooth distribution as shown in FIG. 2 and often shows a more complicated distribution. In such a case, the normalized gain curve obtained by the normalized gain calculation unit 13 of the contrast correction apparatus 10 does not become a smooth upward curve, so that the corrected image may be deteriorated and difficult to see.

For example, when the histogram of the input luminance signal Yi has a plurality of peaks as shown in FIG. 6, the obtained normalized gain curve has a steep slope as shown in FIG. Some portions L ₁ and L ₂ and a portion L ₃ where the change in inclination is abrupt occur.

When the luminance value is corrected based on this normalized gain curve, the conversion based on the portions L ₁ and L ₂ having steep slopes increases the gain, so that the noise component is emphasized and the image quality deteriorates. . Further, in the conversion based on the portion L ₃ where the change in inclination is abrupt, a large distortion is generated in the waveform of the luminance signal straddling the L ₃ , so that the gradation is insufficient (crushing occurs) and the image quality deteriorates. It will be.

  The present invention has been made in view of such a situation, and makes it possible to improve the contrast regardless of the luminance distribution of the input image.

An image processing apparatus according to a first aspect of the present invention is an image processing apparatus that corrects contrast of an image, an acquisition unit that acquires a histogram of luminance signals for one image, and a dynamic range of luminance based on the histogram. A creation means for creating a region frequency distribution indicating the number of pixels belonging to each region by dividing into a predetermined number of regions, and an accumulation assuming a case where luminance signals for one image are evenly distributed in the dynamic range of luminance Calculating means for calculating a deviation at a boundary of the area between a straight line corresponding to the histogram and a curve obtained by combining the straight lines based on the area frequency distribution; and dividing the calculated deviation by the number of gradations of the luminance signal. the normalized using the deviation normalized to the coefficient, Awa set of odd function and even function the curve that combines the line based on the region frequency distribution In approximating the input luminance signal before contrast correction, and generation means for generating a gain function to output a luminance signal after contrast correction, the contrast of the image by correcting on the basis of the luminance signal to the gain function And a correcting means for correcting .

  The image processing apparatus according to the first aspect of the present invention delays and supplies the luminance signal to the correction unit for a time required for processing by the acquisition unit, the generation unit, the calculation unit, and the generation unit. Delay means may further be included.

  The image processing apparatus according to the first aspect of the present invention may further include a mitigation unit that mitigates fluctuations in the time direction of the gain function generated by the generation unit and supplies the fluctuation to the correction unit.

An image processing method according to a first aspect of the present invention is an image processing method of an image processing apparatus that corrects contrast of an image, acquires a histogram of luminance signals for one image, and based on the histogram, a dynamic range of luminance. Is divided into a predetermined number of regions, a region frequency distribution indicating the number of pixels belonging to each region is created, and a cumulative histogram assuming the case where the luminance signal for one image is evenly distributed in the luminance dynamic range A deviation at a boundary of the region between a corresponding straight line and a curve obtained by combining the straight lines based on the region frequency distribution is calculated, and the calculated deviation is normalized by dividing by the number of gradations of the luminance signal. using phased the deviation coefficient was the curve that combines the line based on the region frequency distribution is approximated by a combination of an odd function and even function, Contra Input bets uncorrected luminance signal to generate a gain function to output a luminance signal after contrast correction, comprising the step of correcting the contrast of the image by correcting on the basis of the luminance signal to the gain function.

A program according to the first aspect of the present invention is a program for correcting the contrast of an image, obtains a histogram of luminance signals for one image, and sets a dynamic range of luminance to a predetermined number based on the histogram. A region frequency distribution indicating the number of pixels belonging to each region is created, and a straight line corresponding to a cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance, , Calculating a deviation at the boundary of the region with a curve obtained by combining straight lines based on the region frequency distribution , normalizing the deviation by dividing the calculated deviation by the number of gradations of the luminance signal, and normalizing the deviation the use in coefficient approximating the curve that combines the line based on the region frequency distribution with a combination of an odd function and even function, the contrast uncorrected Enter the degrees signal, executes a process including a step of correcting the contrast of the image to the computer by generating a gain function to output a luminance signal after contrast correction is corrected based on the luminance signal to the gain function Let

In the first aspect of the present invention, based on a histogram of luminance signals for one image, the luminance dynamic range is divided into a predetermined number of regions, and a region frequency distribution indicating the number of pixels belonging to each region is created. Calculates the deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve combining the straight line based on the region frequency distribution is, calculated deviation is normalized by being divided by the number of gradations of the luminance signal, using the normalized deviation coefficient, the curve odd function that combines the linear region based frequency distribution and even functions approximated by a combination of the input luminance signal before contrast correction, it is generated gain function to output a luminance signal after contrast correction, gain the luminance signal The contrast of the image is corrected by being corrected based on the number.

An image processing apparatus according to a second aspect of the present invention is an image processing apparatus that corrects contrast of an image, an acquisition unit that acquires a histogram of a luminance signal for one image, and a dynamic range of luminance based on the histogram. A creation means for creating a region frequency distribution indicating the number of pixels belonging to each region by dividing into a predetermined number of regions, and an accumulation assuming a case where luminance signals for one image are evenly distributed in the dynamic range of luminance Calculating means for calculating a deviation at a boundary of the area between a straight line corresponding to the histogram and a curve obtained by combining the straight lines based on the area frequency distribution; and dividing the calculated deviation by the number of gradations of the luminance signal. the normalized using the deviation normalized to the coefficient, Awa set of odd function and even function the curve that combines the line based on the region frequency distribution In approximating the input luminance signal before contrast correction, and generation means for generating a gain function to output a luminance signal after contrast correction, based on the histogram, the cumulative frequency from the frequency corresponding to the minimum value of the luminance A determination means for determining a luminance having a predetermined value as a black level, a black level compensation function for converting the luminance signal of the black level into a luminance signal of a predetermined pedestal level, and the generated black level compensation function Correction means for correcting the gain function based on the correction function, and correction means for correcting the contrast of the image by correcting the luminance signal based on the gain function corrected by the correction means.

  The image processing apparatus according to the second aspect of the present invention may further include a removing unit that removes a high-frequency component of the luminance signal in a stage preceding the obtaining unit.

An image processing method according to a second aspect of the present invention is an image processing method of an image processing apparatus that corrects the contrast of an image, acquires a histogram of luminance signals for one image, and based on the histogram, a dynamic range of luminance. Is divided into a predetermined number of regions, a region frequency distribution indicating the number of pixels belonging to each region is created, and a cumulative histogram assuming the case where the luminance signal for one image is evenly distributed in the luminance dynamic range A deviation at a boundary of the region between a corresponding straight line and a curve obtained by combining the straight lines based on the region frequency distribution is calculated, and the calculated deviation is normalized by dividing by the number of gradations of the luminance signal. using phased the deviation coefficient was the curve that combines the line based on the region frequency distribution is approximated by a combination of an odd function and even function, Contra Input bets uncorrected luminance signal to generate a gain function to output a luminance signal after contrast correction, based on the histogram, the brightness cumulative frequency from the frequency corresponding to the minimum value of luminance becomes the predetermined value Determining a black level, generating a black level compensation function for converting the black level luminance signal into a predetermined pedestal level luminance signal, correcting the gain function based on the generated black level compensation function, Correcting the contrast of the image by correcting the luminance signal based on the corrected gain function.

A program according to a second aspect of the present invention is a program for correcting the contrast of an image, obtains a histogram of luminance signals for one image, and sets a dynamic range of luminance to a predetermined number based on the histogram. A region frequency distribution indicating the number of pixels belonging to each region is created, and a straight line corresponding to a cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance, , Calculating a deviation at the boundary of the region with a curve obtained by combining straight lines based on the region frequency distribution , normalizing the deviation by dividing the calculated deviation by the number of gradations of the luminance signal, and normalizing the deviation the use in coefficient approximating the curve that combines the line based on the region frequency distribution with a combination of an odd function and even function, the contrast uncorrected Enter the degrees signal to generate a gain function to output a luminance signal after contrast correction, determined on the basis of the histogram, the brightness cumulative frequency from the frequency corresponding to the minimum value of luminance becomes the predetermined value to the black level Generating a black level compensation function for converting the black level luminance signal into a predetermined pedestal level luminance signal, correcting the gain function based on the generated black level compensation function, A computer is caused to execute a process including a step of correcting the contrast of the image by performing correction based on the corrected gain function.

In the second aspect of the present invention, based on the histogram of the luminance signal for one image, the luminance dynamic range is divided into a predetermined number of regions, and a region frequency distribution indicating the number of pixels belonging to each region is created, The deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve combining the straight line based on the region frequency distribution is calculated. , the calculated deviations are normalized by being divided by the number of gradations of the luminance signal, using the normalized deviation coefficient, the curve that combines the linear region based frequency distribution of odd function and even function A gain function approximated by the combination and having the luminance signal before contrast correction as input and the luminance signal after contrast correction as output is generated. Also, based on the generated histogram, the luminance at which the cumulative frequency from the frequency corresponding to the minimum luminance value becomes a predetermined value is determined as a black level, and the black level luminance signal is converted into a predetermined pedestal level luminance signal. A black level compensation function is generated, the gain function is corrected based on the generated black level compensation function, and the brightness signal is corrected based on the corrected gain function, thereby correcting the contrast of the image. The

  According to one aspect of the present invention, the contrast can be improved regardless of the luminance distribution of the input image.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings.

  FIG. 8 shows a first configuration example of a contrast correction apparatus according to an embodiment of the present invention.

  This contrast correction device 20 receives the luminance signal Yi of the image and outputs a corrected (improved) luminance signal Yo, and includes a histogram data acquisition unit 21, a three-band cumulative frequency conversion unit 22, and a deviation calculation unit 23. , A gain function generation unit 24, an LUT luminance conversion unit 25, and a timing control unit 26.

  In the contrast correction device 20, the luminance signal Yi input from the previous stage is input to the histogram data acquisition unit 21 and the LUT luminance conversion unit 25.

  The histogram data acquisition unit 21 counts the histogram data of the luminance signal Yi for one input image (one field or one frame), that is, the frequency of each luminance (pixel appearance frequency), and accumulates three bands. Output to the frequency conversion unit 22.

  FIG. 9 shows a detailed first configuration example of the histogram data acquisition unit 21. In the first configuration example of the histogram data acquisition unit 21, in accordance with control from the timing control unit 26, the value of the luminance signal Yi input from the previous stage is output to the RAM 32 as a read address during the effective period of the luminance signal, and the luminance signal is output. In the invalid period (such as vertical blanking period), the switch 31 that outputs the read address supplied from the timing control unit 26 to the RAM 32, the RAM 32 that outputs the value held in the read address input from the switch 31, and An adder 33 adds 1 to the value output from the RAM 32 and returns it to the RAM 32. The value returned from the adder 33 to the RAM 32 is held at the same address as before the output of the RAM 32.

  A second configuration example of the histogram data acquisition unit 21 will be described later with reference to FIG.

  Returning to FIG. The 3-band cumulative frequency conversion unit 22 divides the dynamic range of luminance into three regions A1, A2, A3 based on the histogram data input from the histogram data acquisition unit 21, and belongs to the frequency regions A1, A2, A3, respectively. The number of pixels having luminance (hereinafter referred to as cumulative frequency in frequency areas A1 to A3) is counted, and the result (hereinafter referred to as 3-band cumulative histogram) is output to deviation calculating unit 23.

  Based on the three-band cumulative histogram input from the three-band cumulative frequency converter 22, the deviation calculating unit 23, for example, as shown in FIG. 10, the origin (0, 0) and (maximum luminance of cumulative frequency area A3, cumulative frequency maximum Value) and the origin (0, 0), (maximum brightness in frequency area A1, cumulative frequency in frequency area A1), (maximum brightness in frequency area A2, cumulative frequency in frequency area A1 + frequency) And a curve g connecting four points of (the cumulative frequency of the region A2) and (the maximum luminance of the frequency region A3, the cumulative frequency of the frequency region A1 + the cumulative frequency of the frequency region A2 + the cumulative frequency of the frequency region A3), The deviation a between the straight line f and the curve g at the maximum luminance of A1 and the deviation b between the straight line f and the curve g at the maximum luminance of the frequency area A2 are calculated, and a + b and a−b are further calculated. Divide by maximum value Normalized by Rukoto outputs normalized a, b, and a + b and a-b to the gain function generating unit 24.

The gain function generation unit 24 converts the curve g that bends at an acute angle into a gain function gc (x) having a smooth slope that is a combination of an even function and an odd function. Specifically, COS (x) is used as the even function, sin (x) is used as the odd function, and the normalized a, b, a + b and a−b input from the deviation calculating unit 23 are used. A gain function gc (x) is generated as in the following equation.
gc (x) = x + ks ・ (ab) ・ sin (2π (x-xp) / (xw-xp))
+ kc ・ (a + b) ・ (1-cos (2π (x-xp) / (xw-xp)))
xp ≦ x ≦ xw
gc (x) = x
0 ≦ x <xp, xw <x ≦ x _max

Here, x is input luminance, and ks and kc are arbitrary coefficients. Also, x _max is the maximum luminance value, xp is the luminance pedestal level, xw is the luminance white peak level, and when the luminance is 8 bits, 0 ≦ x ≦ 255 and x _max = 255. In this case, the pedestal level xp = 16 and the white level xw = 235 are generally used.

  Note that the gain function gc (x) generated by the gain function generator 24 has four patterns of shapes according to the signs of the deviations a and b. That is, the first pattern has a shape based on the -sin function as shown in FIG. The second pattern is a shape based on the + cos function as shown in FIG. 14, for example, based on the histogram shown in FIG. 12 and generating the 3-band cumulative histogram shown in FIG. is there. The third pattern is, for example, a shape based on the -cos function as shown in FIG. 17, which is generated based on the 3-band cumulative histogram shown in FIG. 16 based on the histogram shown in FIG. is there. The fourth pattern is, for example, a shape based on the + sin function as shown in FIG. 20, which is generated based on the 3-band cumulative histogram shown in FIG. 19 based on the histogram shown in FIG. is there.

  Returning to FIG. The gain function generation unit 24 outputs the generated gain function gc (x) to the LUT luminance conversion unit 25. The LUT luminance conversion unit 25 holds the gain function gc (x) input from the gain function generation unit 24 as a lookup table (hereinafter referred to as LUT), and the luminance signal Yi input from the previous stage of the contrast correction device 20. Is converted into a luminance signal Yo using the holding LUT and output to the subsequent stage. The timing control unit 15 controls the operation timing of the histogram data acquisition unit 21 to the LUT luminance conversion unit 25.

  FIG. 21 shows a detailed configuration example of the LUT luminance conversion unit 25. The LUT luminance conversion unit 25 includes a switch 41 and an LUT 42. In accordance with control from the timing control unit 26, the switch 41 outputs the value of the luminance signal Yi input from the previous stage to the LUT 21 as a read address during the effective period of the luminance signal. In addition, the write address supplied from the timing control unit 26 is output to the LUT 42 during the invalid period (such as the vertical blanking period) of the luminance signal. The LUT 42 is configured by, for example, a RAM, and stores a value obtained by multiplying the gain function gc (x) supplied from the gain function generation unit 24 by the maximum luminance in accordance with a write address input via the switch 41 as a lookup table. The LUT 42 is configured to output a value held at a read address input from the switch 41 (that is, a corrected luminance signal Yo).

  Next, the operation of the contrast correction apparatus 20 (hereinafter referred to as contrast correction processing) will be described with reference to the flowchart of FIG.

  When the luminance signal Yi of the image is input from the previous stage to the contrast correction device 20, in step S1, the histogram data acquisition unit 21 outputs the luminance signal for one input image (for one field or one frame). Histogram data corresponding to Yi is generated and output to the 3-band cumulative frequency converter 22. In step S <b> 2, the 3-band cumulative frequency conversion unit 22 generates a 3-band cumulative histogram based on the histogram data input from the histogram data acquisition unit 21 and outputs the 3-band cumulative histogram to the deviation calculation unit 23.

  In step S <b> 3, the deviation calculation unit 23 calculates normalized a, b, a + b, and a−b based on the three-band cumulative histogram input from the three-band cumulative frequency conversion unit 22, and sends it to the gain function generation unit 24. Output. In step S4, the gain function generator 24 generates a gain function gc (x) based on the normalized a, b, a + b, and a−b input from the deviation calculator 23, and performs LUT luminance conversion. To the unit 25.

  In step S <b> 5, the LUT luminance conversion unit 25 holds the gain function gc (x) input from the gain function generation unit 24 as a lookup table. In step S6, the LUT luminance conversion unit 25 converts the luminance signal Yi input from the previous stage into a luminance signal Yo using the held LUT, and outputs it to the subsequent stage.

  According to the contrast correction processing of the contrast correction device 20 described above, even if a plurality of peaks appear in the histogram as shown in FIG. 6, for example, the three-band accumulation shown in FIG. A histogram is generated, a gain function gc (x) having the shape of a -sin function including a smooth curve shown in FIG. 24 is generated, and the luminance is corrected by the generated gain function gc (x), so that it is output. The luminance signal Yo that has the noise component is emphasized and the contrast is appropriately improved without deterioration of the image quality or deterioration of the image quality due to insufficient gradation.

  Note that the processing in steps S1 to S5 is performed during a period in which the luminance signal Yi for one image is input. Therefore, the luminance signal Yi corrected in the LUT luminance conversion unit 25 and the LUT used at that time are 1 Although there will be a shift of the image, normally there is no great difference in the luminance characteristics of the preceding and following images, so this shift of one image does not cause a problem such as deterioration of the corrected image.

  Above, description of the contrast correction process by the contrast correction apparatus 20 is complete | finished.

  Next, a configuration example of a contrast correction apparatus (hereinafter referred to as a second example) that can cope with a problem (for example, a scene change scene) of a deviation of one image between the luminance signal Yi to be corrected and the LUT used at that time. Will be described.

  FIG. 25 shows a second configuration example of the contrast correction apparatus. This contrast correction device 50 is obtained by adding a delay unit 51 to the preceding stage of the LUT luminance conversion unit 25 with respect to the contrast correction device 20 which is the first configuration example shown in FIG. Is the same as that of the contrast correction apparatus 20, and is given the same reference numerals.

  The delay unit 51 delays the luminance signal Yi input from the previous stage of the contrast correction device 50 by one image period and outputs the delayed signal to the LUT luminance conversion unit 25. Accordingly, in the LUT luminance conversion unit 25 of the contrast correction device 50, the luminance signal Yi to be corrected and the LUT used at that time are not shifted by one image. It becomes possible to correct.

  Next, a configuration example (hereinafter referred to as a third configuration example) of the contrast correction device corresponding to a case where the change between the images before and after the luminance signal Yo sequentially output from the contrast correction device 20 is not sufficiently relaxed. explain.

  FIG. 26 shows a third configuration example of the contrast correction apparatus. This contrast correction device 60 is obtained by adding an IIR filter 61 between the gain function generation unit 24 and the LUT luminance conversion unit 25 to the contrast correction device 20 which is the first configuration example shown in FIG. The other components are the same as those of the contrast correction apparatus 20, and are given the same reference numerals.

  An IIR (Infinite Impulse Response) filter 61 receives the gain function gc (x) from the gain function generation unit 24 and outputs a filtering result to the LUT luminance conversion unit 25. Therefore, since the fluctuation in the time axis direction of the gain function gc (x) supplied to the LUT luminance conversion unit 25 is suppressed, the LUT luminance conversion unit 25 can perform luminance conversion that is more stable, that is, the change is moderate. Become.

  The IIR filter 61 is provided with a function for detecting the scene change based on the difference between the histograms of the preceding and following images and changing the time constant setting of the IIR filter according to the detection result to improve the response speed. Also good.

  Further, the delay unit 51 may be added to the contrast correction device 60 as well as the contrast correction device 50.

  Next, a configuration example (hereinafter referred to as a fourth configuration example) of a contrast correction apparatus that improves the pixel value by correcting not only the contrast by the gain control characteristic but also the black level of the luminance signal will be described. . The black level correction means that, for example, in a television broadcast signal, the transmitted black level fluctuates due to differences in shooting conditions, recording equipment, etc., but this black level is received on the receiving side (for example, a television receiver). ) Means that the image quality is improved by matching the pedestal level which is the black level.

  FIG. 27 shows a fourth configuration example of the contrast correction apparatus. The contrast correction apparatus 70 is different from the contrast correction apparatus 20 of the first configuration example shown in FIG. 8 in that it includes a low-pass filter 71, a black level detection unit 72, a black level compensation gain calculation unit 73, and an IIR filter 74. The other components are the same as those of the contrast correction apparatus 20 and are given the same reference numerals.

The low-pass filter 71 provided in the previous stage of the histogram data acquisition unit 21 receives the luminance signal Yi from the previous stage of the contrast correction device 70, removes the high-frequency component, and outputs it to the histogram data acquisition unit 21. Since the high-frequency component is removed by the low-pass filter 71, fluctuations in the vicinity of the black level of the luminance signal Yi are suppressed. The low-pass filter 71 may have a relatively simple configuration as shown in the following equation, for example.
flpf (Z) = (1 + 2 · Z ⁻¹ + Z ⁻² ) / 4
However, Z ⁻¹ represents a delay of one clock cycle.

  The black level detection unit 72 provided at the subsequent stage of the histogram data acquisition unit 21 receives the histogram data from the histogram data acquisition unit 21, and the cumulative frequency from luminance = 0 is a predetermined value (for example, 3% of the total number of pixels). The black level BL is determined as the luminance that has reached the value of (5), and the black level BL is output to the black level compensation gain calculator 73.

  A black level compensation gain calculation unit 73 provided at the subsequent stage of the gain function generation unit 24 calculates a black compensation amount ka corresponding to the black level BL input from the black level detection unit 72, and gain based on the black compensation amount ka. The gain function gc (x) input from the function generation unit 24 is corrected, and the corrected gain function gct (x) obtained as a result is output to the IIR filter 74.

  The IIR filter 74 receives the gain function gct (x) from the black level compensation gain calculator 73 and outputs the filtering result to the LUT luminance converter 25.

  Next, the operation of the contrast correction apparatus 70 (hereinafter referred to as contrast correction processing) will be described with reference to the flowchart of FIG. 28, taking as an example the case where the luminance signal Yi of the histogram shown in FIG. 6 is input. . The contrast correction process including the black level correction is effective when the frequency in a low luminance range is small and so-called black floating of the image occurs as in the histogram shown in FIG.

  When the luminance signal Yi of the image is input from the previous stage to the contrast correction device 70, in step S11, the low-pass filter 71 performs the filtering of the luminance signal Yi, and the luminance signal Yi from which the high frequency component obtained as a result is removed. Is output to the histogram data acquisition unit 21. In step 12, the histogram data acquisition unit 21 generates histogram data corresponding to the luminance signal Yi for one input image (one field or one frame) to generate the 3-band cumulative frequency conversion unit 22 and the black level. Output to the detector 72.

  In step S <b> 13, the 3-band cumulative frequency conversion unit 22 generates a 3-band cumulative histogram based on the histogram data input from the histogram data acquisition unit 21 and outputs the 3-band cumulative histogram to the deviation calculation unit 23. In the case of the present example, the 3-band cumulative histogram shown in FIG. 23 is output.

  In step S <b> 14, the deviation calculation unit 23 calculates normalized a, b, a + b, and a−b based on the three-band cumulative histogram input from the three-band cumulative frequency conversion unit 22, and sends it to the gain function generation unit 24. Output. In step S15, the gain function generator 24 generates a gain function gc (x) based on the normalized a, b, a + b, and a−b input from the deviation calculator 23, and performs LUT luminance conversion. To the unit 25. In this example, the gain function gc (x) shown in FIG. 24 is output.

  In step S <b> 16, the black level detection unit 72 has a cumulative frequency from luminance = 0 reaches a predetermined value (for example, a value of 3% of the total number of pixels) based on the histogram data input from the histogram data acquisition unit 72. The determined luminance is determined as the black level BL, and the black level BL is output to the black level compensation gain calculator 73. In the present example, it is assumed that luminance = 84 is determined as the black level BL.

  Note that the processes in steps S13 to S15 and the process in step S16 are actually executed in parallel.

  In step S17, the black level compensation gain calculation unit 73 performs black expansion which is the difference between the gain function gc (BL) × 255 at the black level BL shown in FIG. 29 and the gain function gc (xp) × 255 at the pedestal level xp. The amount ΔB is calculated. In this example, it is assumed that 46 is calculated as the black extension amount ΔB.

In step S18, as shown in FIG. 30, the black level compensation gain calculation unit 73 sets a bending point CP near the middle of the value that the luminance can take, has a slope of 1 in the range beyond the bending point CP, and is bent. A gain control characteristic fauto (x) is calculated which is smoothly connected at the point CP and is a quadratic function of the luminance signal Yi in the range below the bending point CP and can draw the luminance signal Yi of the black level BL to the pedestal level.
fauto (x) = x−ka ・ (x−CP) ² 0 ≦ x ≦ CP
fauto (x) = x CP ≦ x ≦ 255

Further, the black level compensation amount ka is calculated by substituting the relationship at the black level BL into the above equation.
BL−ΔB = BL−ka ・ (BL−CP) ²
ΔB = ka · (BL−CP) ²
ka = ΔB / (BL-CP) ²
In the present example, the black level compensation amount ka = 46 / (84−CP) ² is calculated.

In step S19, the black level compensation gain calculation unit 73 corrects the gain function gc (x) below the bending point CP based on the black level compensation amount ka as follows, and the corrected gain function gct (x ) And output to the LUT luminance conversion unit 25 via the IIR filter 74.
gct (x) = x−ka · (x−CP) ² 0 ≦ x <xp
gct (x) = x + ks ・ (ab) ・ sin (2π (x-xp) / (xw-xp))
+ Kc ・ (a + b) ・ (1-cos (2π (x-xp) / (xw-xp)))
−ka ・ (x−CP) ² xp ≦ x ≦ CP
gct (x) = x + ks ・ (ab) ・ sin (2π (x-xp) / (xw-xp))
+ Kc ・ (a + b) ・ (1-cos (2π (x-xp) / (xw-xp))) CP <x ≦ xw
gct (x) = x xw <x ≦ x _max

  In this example, the corrected gain function gct (x) is as shown in FIG. That is, the luminance signal Yi of the black level BL is drawn to the pedestal level and has contrast improvement characteristics.

  In step S20, the LUT luminance conversion unit 25 holds the gain function gct (x) input via the IIR filter 74 as a lookup table. In step S21, the LUT luminance conversion unit 25 converts the luminance signal Yi input from the previous stage into the luminance signal Yo using the held LUT, and outputs it to the subsequent stage.

  According to the contrast correction processing of the contrast correction apparatus 70 described above, stable black level reproduction can be achieved by contrast improvement and black level compensation. In particular, in the contrast correction device 70, the IIR filter 74 can share the time constant related to the contrast improvement and the relaxation of the black level compensation fluctuation, so that a more stable time response characteristic can be obtained. Accordingly, the coefficient of contrast improvement amount and the black level compensation amount can be set larger, and the image quality improvement effect can be further extracted.

  FIG. 32 shows histograms of the luminance signal Yo after conversion when converted with the gain function gc (x) and when converted with the gain function gct (x). In the case of conversion with the gain function gc (x), the frequency appears to be distributed over the entire luminance range, but there is almost no frequency with a luminance of 50 or less, and the state is floated black. On the other hand, when converted by the gain function gct (x), the frequencies are distributed over the entire region including the range where the luminance is 50 or less. Therefore, it can be seen that the luminance dynamic range is sufficiently used to improve the contrast.

  Above, description of the contrast correction process by the contrast correction apparatus 70 is complete | finished.

Incidentally, for example, in the case of an image having a uniform background such as the image shown in FIG. 33, the histogram shows the luminance signal corresponding to the subject with a frequency corresponding to the background (around 150) as shown in FIG. Often higher than the frequency. When contrast correction is performed based on such a histogram, the contrast of the portion of the subject having the same luminance as the background luminance is unnecessarily corrected, resulting in an unnatural and unsightly image.

  In order to deal with such a background problem, a second configuration example of the histogram data acquisition unit 21 is proposed. FIG. 35 illustrates a second configuration example of the histogram data acquisition unit 21. The second configuration example of the histogram data acquisition unit 21 is different from the first configuration example shown in FIG. 9 in that the background determination unit 101, the inverter 102, the background image count unit 103, the averaging unit 104, and the addition The part 105 is added, and the other components are the same as those in the first configuration example, and are given the same reference numerals.

  The background determination unit 101 determines whether or not the luminance signal Yi input from the previous stage corresponds to a background pixel. When it is determined that the luminance signal Yi corresponds to a background pixel, When it is determined that the pixel does not correspond to the pixel, 0 is output to the inverter 102 and the background pixel count unit 103.

  FIG. 36 shows a detailed configuration example of the background determination unit 101. The background determination unit 101 includes a low-pass filter 111, a band-pass filter 112, an absolute value conversion unit 113, a low-pass filter 114, a comparison unit 115, and a continuity determination unit 116.

  The low-pass filter 111 removes the high-frequency component of the luminance signal Yi input from the previous stage and outputs it to the band-pass filter 112. The band pass filter 112 passes only a predetermined frequency band in the output of the low pass filter 111 to the absolute value converter 113.

  Two examples of frequency characteristics of the low-pass filter 111 and the band-pass filter 112 are shown in FIG. In the figure, the horizontal axis indicates the frequency, and the vertical axis indicates the gain. Fs is a sampling frequency, and fs / 2 is a maximum passing frequency of a signal sampled by fs. As shown in the figure, the overall characteristic of the low-pass filter 111 and the band-pass filter 112 is that the low-pass side of the pass band of the low-pass band-pass filter 112 of the signal pass band passes through the high-pass side. To be Thereby, the high frequency noise component of the luminance signal Yi is removed.

  Such a comprehensive characteristic is that, particularly when the luminance input signal Yi is a signal using a moving image compression technique such as MPEG, the amount of high frequency noise components may vary for each image. However, it is possible to prevent the gain characteristic from fluctuating due to fluctuation of the determination frequency of the background portion in the fluctuation cycle.

  Returning to FIG. The absolute value conversion unit 113 outputs the absolute value of the input from the band pass filter 112 to the low pass filter 114. The low-pass filter 114 attenuates and smoothes the higher-order frequency generated by the absolute value conversion of the absolute value conversion unit 113 and outputs it to the comparison unit 115. The comparison unit 115 compares the output of the low-pass filter 114 with a predetermined determination level. When the output of the low-pass filter unit 114 is equal to or higher than the predetermined determination level, 1 is output, and the output of the low-pass filter unit 114 is the predetermined determination level. If it is smaller than the level, 0 is output to the continuity determination unit 116.

  As shown in FIG. 38, the continuity determination unit 116 includes delay units (D) 121 to 124 that output a 1-bit input delayed by one sample clock, and a NOR operation unit 125. The comparison unit 115 When the output of 5 is 0 continuously, it is determined as the background and 1 is output to the subsequent stage, and 0 is output to the subsequent stage at other times. It should be noted that by increasing the number of delay portions, it is possible to detect a portion with a more flat luminance change.

  Returning to FIG. Inverter 102 inverts the input from background determination unit 101 and outputs the result to addition unit 33. Therefore, the frequency held in the RAM 32 is incremented only when the pixel does not correspond to the background pixel. The background pixel counting unit 103 sequentially adds the inputs from the background determination unit 101 and outputs the added value (background counter value) to the averaging unit 104 when the luminance signal Yi for one image is input. Reset the background counter value to zero.

  The averaging unit 104 divides the background counter value input from the background pixel counting unit 103 by the number of gradations of the dynamic range of luminance (256 in the case of 8 bits), and outputs the quotient to the adding unit 105. This division is easily calculated by, for example, truncating the output of the background pixel count unit 103 from the lower bits by the number of bits of luminance. For example, when the luminance is 8 bits, the lower 8 bits of the output of the background pixel count unit 103 may be discarded.

  The adding unit 105 adds the value input from the averaging unit 104 to the frequency of each luminance reflecting the portion other than the background output from the RAM 32 for each image, and outputs the result to the subsequent stage. Therefore, the total frequency of the histogram output to the subsequent stage from the adding unit 105 is the same as the number of pixels for one image.

  Next, the operation of the second configuration example of the histogram data acquisition unit 21 (hereinafter referred to as a histogram generation process with the background removed) will be described with reference to the flowchart of FIG.

  In step S31, the background determination unit 101 pays attention to one pixel of the luminance signal Yi sequentially input from the previous stage, and determines whether or not the pixel belongs to the background in step S32.

  If it is determined in step S32 that the pixel belongs to the background, the process proceeds to step S33. In step S <b> 33, the background determination unit 101 outputs 1 to the inverter 102 and the background count unit 103. The 1 input to the inverter 102 is inverted to 0 and supplied to the adder 33. 1 input to the background pixel count unit 103 is added to the background counter value held so far (that is, the background counter value is incremented by 1). Therefore, the frequency of the pixel is not reflected in the RAM 32 but only in the background counter value.

  Conversely, if it is determined in step S32 that the pixel does not belong to the background, the process proceeds to step S34. In step S <b> 34, the background determination unit 101 outputs 0 to the inverter 102 and the background count unit 103. The 0 input to the inverter 102 is inverted to 1 and supplied to the adder 33. The 0 input to the background pixel count unit 103 is added to the background counter value held so far. Therefore, the frequency of the pixel is reflected in the RAM 32 (that is, the frequency corresponding to the luminance signal Yi of the pixel is incremented by 1) and is not substantially reflected in the background counter value.

  In step S35, the background determination unit 101 determines whether or not attention is paid to one pixel of the luminance signal Yi sequentially input from the previous stage. Return to and repeat the subsequent processing. If it is determined in step S35 that attention is paid to pixels for one image, the process proceeds to step S36.

  In step S <b> 36, the background pixel count unit 103 outputs the background counter value to the averaging unit 104. The averaging unit 104 divides the background counter value input from the background pixel counting unit 103 by the number of gradations of the luminance dynamic range, and outputs the quotient to the adding unit 105.

  In step S <b> 37, the RAM 32 outputs the value held so far in accordance with the read address designated from the timing control unit 26 via the switch 31, thereby adding the frequency of each luminance for one image excluding the background. Output to the unit 105. The adding unit 105 adds the value input from the averaging unit 104 to the frequency of each luminance reflecting the portion other than the background output from the RAM 32 for each image, and outputs the result to the subsequent stage.

  As shown in FIG. 40, the histogram output from the adding unit 105 has a uniform background as shown in FIG. 40 even if the image input to the histogram data acquiring unit 21 has a uniform background as shown in FIG. The luminance frequency corresponding to the background is evenly distributed over the entire luminance range without protruding the luminance frequency corresponding to. Therefore, if contrast correction is performed based on this histogram, it is possible to prevent an image from being unnatural and unsightly because the contrast of the portion of the subject having the same luminance as the background is unnecessarily corrected. be able to.

  As described above, the method of generating the histogram by removing the background has the effect of removing the frequency of the flat portion not only for a natural image but also for an image having many flat portions such as a computer graphic or an animation image. is there. Furthermore, test signals such as color bars also have a lot of flat parts in the same way, and unnecessary gain characteristics were generated in the conventional example, but they are also effective for such signals. be able to.

  This is the end of the description of the histogram generation process in which the background is removed according to the second configuration example of the histogram data acquisition unit 21.

  By the way, the series of processes described above can be executed by hardware, but can also be executed by software. When a series of processing is executed by software, a program constituting the software may execute various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed from a recording medium in a general-purpose personal computer configured as shown in FIG.

The personal computer 200 incorporates a CPU (Central Processing Unit) 201. An input / output interface 205 is connected to the CPU 201 via the bus 204. A ROM (Read Only Memory) 202 and a RAM (Random Access Memory) 203 are connected to the bus 204.

  An input / output interface 205 includes a keyboard for inputting operation commands by a user, an input unit 206 including an input device such as a mouse, a CRT (Cathode Ray Tube) or an LCD (Liquid Crystal Display) for displaying an image to be processed, etc. Output unit 207 consisting of a display, a storage unit 208 consisting of a hard disk drive for storing programs and various data, a modem, a LAN (Local Area Network) adapter, etc., and communication processing via a network represented by the Internet. A communication unit 209 to be executed is connected. Also, a magnetic disk (including a flexible disk), an optical disk (including a CD-ROM (Compact Disc-Read Only Memory), a DVD (Digital Versatile Disc)), a magneto-optical disk (including an MD (Mini Disc)), or a semiconductor A drive 210 for reading / writing data from / to a recording medium 211 such as a memory is connected.

  A program for causing the personal computer 200 to execute the above-described series of processing is supplied to the personal computer 200 while being stored in the recording medium 211, read by the drive 210, and installed in a hard disk drive built in the storage unit 208. Has been. The program installed in the storage unit 208 is loaded from the storage unit 208 to the RAM 203 and executed in response to a command from the CPU 201 corresponding to a command from the user input to the input unit 206.

  In this specification, the steps executed based on the program are executed in parallel or individually even if they are not necessarily processed in time series, as well as processes executed in time series according to the described order. It also includes processing.

  The program may be processed by a single computer, or may be distributedly processed by a plurality of computers. Furthermore, the program may be transferred to a remote computer and executed.

  Further, in this specification, the system represents the entire apparatus constituted by a plurality of apparatuses.

  The embodiment of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the gist of the present invention.

It is a block diagram which shows an example of a structure of the conventional contrast correction apparatus. It is a figure which shows an example of the histogram of a brightness | luminance. FIG. 3 is a diagram illustrating a cumulative histogram corresponding to the histogram of FIG. 2. It is a figure which shows the gain control characteristic based on the accumulation histogram of FIG. It is a figure which shows the histogram after the correction | amendment corresponding to the histogram of FIG. It is a figure which shows an example of the histogram of a brightness | luminance. It is a figure which shows the gain control characteristic corresponding to the histogram of FIG. It is a block diagram which shows the 1st structural example of the contrast correction apparatus to which this invention is applied. It is a block diagram which shows the 1st structural example of the histogram data acquisition part of FIG. It is a figure which shows an example of 3 band cumulative histogram. It is a figure which shows the gain function based on the 3 band cumulative histogram of FIG. It is a figure which shows an example of the histogram of a brightness | luminance. FIG. 13 is a diagram showing a 3-band cumulative histogram corresponding to the histogram of FIG. 12. It is a figure which shows the gain function based on the 3 band cumulative histogram of FIG. It is a figure which shows an example of the histogram of a brightness | luminance. FIG. 16 is a diagram showing a 3-band cumulative histogram corresponding to the histogram of FIG. 15. It is a figure which shows the gain function based on the 3 band cumulative histogram of FIG. It is a figure which shows an example of the histogram of a brightness | luminance. FIG. 19 is a diagram showing a 3-band cumulative histogram corresponding to the histogram of FIG. 18. FIG. 20 is a diagram illustrating a gain function based on the 3-band cumulative histogram of FIG. 19. It is a block diagram which shows the structural example of the LUT brightness | luminance conversion part of FIG. It is a flowchart explaining the contrast correction process by the 1st structural example of a contrast correction apparatus. FIG. 7 is a diagram showing a 3-band cumulative histogram corresponding to the histogram of FIG. 6. It is a figure which shows the gain function corresponding to the 3 band cumulative histogram of FIG. It is a block diagram which shows the 2nd structural example of the contrast correction apparatus to which this invention is applied. It is a block diagram which shows the 3rd structural example of the contrast correction apparatus to which this invention is applied. It is a block diagram which shows the 4th structural example of the contrast correction apparatus to which this invention is applied. It is a flowchart explaining the operation | movement by the 4th structural example of a contrast correction apparatus. It is a figure which shows an example of the gain function without black level compensation. It is a figure for demonstrating calculation of the coefficient ka. It is a figure which shows the gain function correct | amended with black level compensation corresponding to the gain function of FIG. It is a figure which shows the histogram after correction | amendment by the 4th structural example of a contrast correction apparatus. It is a figure which shows an example of the image containing a uniform background. It is a figure which shows the histogram of the brightness | luminance corresponding to the image of FIG. It is a block diagram which shows the 2nd structural example of a histogram data acquisition part. FIG. 36 is a block diagram illustrating a configuration example of a background determination unit in FIG. 35. It is a figure which shows the characteristic of the low pass filter and band pass filter which comprise a background determination part. It is a block diagram which shows the structural example of the continuity determination part of FIG. It is a flowchart explaining operation | movement of the 2nd structural example of a histogram data acquisition part. It is a figure which shows the histogram from which the background part was output output from the 2nd structural example of a histogram data acquisition part. It is a block diagram which shows the structural example of a general purpose personal computer.

Explanation of symbols

  20 contrast correction device, 21 histogram data acquisition unit, 22 3-band cumulative frequency conversion unit, 23 deviation calculation unit, 24 gain function generation unit, 25 LUT luminance conversion unit, 51 delay unit, 61 IIR filter, 71 low pass filter, 72 black Level detection unit, 73 black level compensation gain calculation unit, 74 IIR filter, 101 background determination unit, 102 inverter, 103 background pixel count unit, 104 averaging unit, 105 addition unit, 201 CPU, 211 recording medium

Claims (9)

In an image processing apparatus for correcting the contrast of an image,
Obtaining means for obtaining a histogram of luminance signals for one image;
Based on the histogram, creating means for dividing a dynamic range of luminance into a predetermined number of regions and creating a region frequency distribution indicating the number of pixels belonging to each region;
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution Calculating means for calculating
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generating means for generating a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
An image processing apparatus comprising: correction means for correcting the contrast of the image by correcting the luminance signal based on the gain function. The image processing apparatus according to claim 1, further comprising a delay unit that delays the luminance signal and supplies the luminance signal to the correction unit for a time required for processing of the acquisition unit, the generation unit, the calculation unit, and the generation unit. . The image processing apparatus according to claim 1, further comprising a relaxation unit that relaxes a time-direction variation of the gain function generated by the generation unit and supplies the gain function to the correction unit. In the image processing method of the image processing apparatus for correcting the contrast of the image,
Obtain a histogram of the luminance signal for one image,
Based on the histogram, the luminance dynamic range is divided into a predetermined number of regions, creating a region frequency distribution indicating the number of pixels belonging to each region,
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution To calculate
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generate a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
An image processing method comprising: correcting the contrast of the image by correcting the luminance signal based on the gain function. A program for correcting the contrast of an image,
Obtain a histogram of the luminance signal for one image,
Based on the histogram, the luminance dynamic range is divided into a predetermined number of regions, creating a region frequency distribution indicating the number of pixels belonging to each region,
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution To calculate
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generate a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
A program that causes a computer to execute processing including a step of correcting the contrast of the image by correcting the luminance signal based on the gain function. In an image processing apparatus for correcting the contrast of an image,
Obtaining means for obtaining a histogram of luminance signals for one image;
Based on the histogram, creating means for dividing a dynamic range of luminance into a predetermined number of regions and creating a region frequency distribution indicating the number of pixels belonging to each region;
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution Calculating means for calculating
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generating means for generating a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
Determining means for determining, on the basis of the histogram, a luminance at which a cumulative frequency from a frequency corresponding to a minimum luminance value is a predetermined value, as a black level;
Correcting means for generating a black level compensation function for converting the black level luminance signal into a predetermined pedestal level luminance signal, and correcting the gain function based on the generated black level compensation function;
An image processing apparatus comprising: correction means for correcting the contrast of the image by correcting the luminance signal based on the gain function corrected by the correction means. The image processing apparatus according to claim 6, further comprising a removing unit that removes a high-frequency component of the luminance signal before the obtaining unit. In the image processing method of the image processing apparatus for correcting the contrast of the image,
Obtain a histogram of the luminance signal for one image,
Based on the histogram, the luminance dynamic range is divided into a predetermined number of regions, creating a region frequency distribution indicating the number of pixels belonging to each region,
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution To calculate
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generate a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
Based on the histogram, determine the luminance at which the cumulative frequency from the frequency corresponding to the minimum luminance value is a predetermined value as the black level,
Generating a black level compensation function for converting the black level luminance signal into a predetermined pedestal level luminance signal, and correcting the gain function based on the generated black level compensation function;
An image processing method comprising: correcting the contrast of the image by correcting the luminance signal based on the corrected gain function. A program for correcting the contrast of an image,
Obtain a histogram of the luminance signal for one image,
Based on the histogram, the luminance dynamic range is divided into a predetermined number of regions, creating a region frequency distribution indicating the number of pixels belonging to each region,
Deviation at the boundary of the region between the straight line corresponding to the cumulative histogram assuming that the luminance signal for one image is evenly distributed in the dynamic range of luminance and the curve obtained by combining the straight line based on the region frequency distribution To calculate
The calculated the difference was normalized by dividing by the number of gradations of the luminance signal, using said deviation normalized to the coefficient, the curve that combines the line based on the region the frequency distribution of the odd function and even function in combination Generate a gain function that receives the luminance signal before contrast correction and outputs the luminance signal after contrast correction , approximated by
Based on the histogram, determine the luminance at which the cumulative frequency from the frequency corresponding to the minimum luminance value is a predetermined value as the black level,
Generating a black level compensation function for converting the black level luminance signal into a predetermined pedestal level luminance signal, and correcting the gain function based on the generated black level compensation function;
A program that causes a computer to execute processing including a step of correcting the contrast of the image by correcting the luminance signal based on the corrected gain function.

Images