JP2015142260A - Image processing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce a depth range while keeping depth distribution within a subject even in the case of an image which has almost uniform existence frequency of a depth.SOLUTION: An image processing system which reduces the depth range of a depth image comprises: an area division unit which divides the depth image into two areas; and a depth change unit which changes the depth of at least one of the two areas by a predetermined amount.

Description

  The present invention relates to an image processing technique for processing depth information.

  For example, in the technology used to display a stereoscopic image by an image display device, stereoscopic display is realized by presenting different images to the left and right eyes of a human, and the displacement of an object in the left and right eye image The depth is reproduced by the parallax. The reproducible depth range depends on the specifications of the display device. When the depth range of an image input to the display device is wider than the reproducible depth range, the depth range needs to be converted before being displayed.

  Patent Document 1 discloses a map conversion method for converting a depth map using the presence frequency of depth in an image so that a better sense of depth can be obtained within a reproducible depth range. A depth map is an image in which a depth value is held in each pixel.

  FIG. 6 is an example of a histogram showing the presence frequency (frequency) of depth. FIG. 6A is a diagram illustrating an example of a histogram indicating the presence frequency of the depth for each pixel of the input image. The vertical axis is frequency and the horizontal axis is class value. FIG. 6A is a bimodal graph, and there is a range of depth where the frequency is zero between the two mountains. For example, such a histogram is obtained when there are two subjects in the image and one is located in front and the other is located in the back. In the following description, the depth value is larger as the subject is closer to the front, and the depth value is smaller as the subject is farther away. The right mountain (peak) is the value for the subject in the foreground, and the left mountain (peak) is the value for the subject in the back. The difference in the class value between the tails of the peaks corresponding to the depth range is W1.

  FIG. 6B is a histogram showing the presence frequency of the depth after the depth conversion is performed on the histogram of FIG. The minimum value of the depth in FIG. 6B is the same as that in FIG. 6A, but the maximum value is smaller than that in FIG. 6A. The position of the left mountain is the same as A, but the maximum value is reduced by shifting the right mountain to the left and attaching the two peaks without any gaps. That is, of the two subjects, the position of the subject in the foreground is moved to the back side while keeping the subject in the back as it is. By converting in this way, the entire depth range can be reduced while maintaining the depth change inside each subject (W1> W2).

  In contrast to depth conversion, tone mapping that converts the range of luminance values of an image also uses a conversion method that uses the frequency of luminance values for each pixel in the image. Tone mapping is a process of converting an image with a wide dynamic range into an image with a narrow dynamic range. Patent Document 2 discloses a conversion method using a mapping function generated using a histogram of luminance values.

JP 2012-134881 A JP 2007-310887 A

  However, in the depth conversion process, in the case of a depth map in which the frequency of the depth is substantially uniform, it is difficult to efficiently reduce the depth range by the conventional depth conversion method using the histogram. FIG. 6C is an example of a histogram when the depth existence frequency of the input image is substantially uniform.

  Also, the tone mapping process that converts the range of the brightness value of the image has the same problem as the depth conversion process. If the frequency of the brightness value of the input image is almost uniform, the conventional technique using a histogram is used. With this conversion method, it has been difficult to efficiently reduce the range of luminance values.

  In this specification, numerical values assigned to each pixel of the image, such as the depth value of each pixel constituting the depth map and the luminance value of each pixel constituting the luminance image, are collectively referred to as depth. In terms of the histogram, the value of the class value corresponds to the depth. A depth image is an image having depth as a parameter.

  The present invention has been made in view of the above-described actual situation, and an object of the present invention is to maintain the depth distribution in the subject even in the case of an image having a substantially uniform depth existence frequency (frequency). An object of the present invention is to provide an image processing method capable of reducing the depth range.

  The present invention is an image processing apparatus for reducing a depth range of a depth image, and an area dividing unit that divides the depth image into two areas, and a depth is changed by a predetermined amount in at least one of the two areas. And a depth changing unit.

  The region dividing unit is a cut surface image generating unit that converts the depth image to generate a cut surface image having a continuous value between adjacent pixels, and corresponds the pixel values of the depth image and the cut surface image. The image processing apparatus includes a cutting processing unit that divides the depth image into two regions by comparing each pixel and extracting a pixel having a pixel value of the depth image larger than the pixel value of the cut surface image.

  Furthermore, it has a depth structure map generation unit that generates a three-dimensional depth structure map from the depth image, and the cut surface image generation unit has a depth value that minimizes the sum of the values of the depth structure map in all pixels. It is characterized by calculating | requiring.

  According to the present invention, it is possible to provide an image processing technique capable of reducing the depth range while maintaining the depth distribution in the subject even in the case of an image having a substantially uniform depth existence frequency.

1 is a functional block diagram illustrating a configuration example of an image processing apparatus according to a first embodiment of the present invention. It is a figure explaining a cut surface. It is a figure explaining the outline | summary of Formula (5). It is a flowchart which shows the image processing method which concerns on the 1st Embodiment of this invention. It is a functional block diagram which shows the example of 1 structure of the image processing apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the image processing method which concerns on the 2nd Embodiment of this invention. It is a figure which shows the example of the histogram which shows the presence frequency of depth.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

(First embodiment)
FIG. 1 is a block diagram illustrating a configuration example of an image processing apparatus according to the first embodiment of the present invention. The image processing apparatus 10 of the present embodiment inputs a depth map, performs depth conversion processing, and outputs a converted depth map. As illustrated in FIG. 1, the image processing apparatus 10 includes a depth map input unit 11, a depth conversion amount calculation unit 12, a depth conversion unit 13, and an output unit 14.

  Details of each unit in the image processing apparatus 10 will be described below. The depth map input unit 11 inputs a depth map and outputs it to the depth conversion amount calculation unit 12. The depth map value input here may be a distance from the camera to the subject obtained by a ranging camera that can acquire the depth for each pixel, or a parallax generated by image processing from an image captured by a stereo camera, or It may be the distance between the stereo camera and the subject calculated from the parallax.

  First, the depth conversion amount calculation unit 12 calculates the minimum value and the maximum value of the depth from the depth map, and sets them as the input depth minimum value pmin and the input depth maximum value pmax, respectively. Next, the depth conversion amount k is calculated from the maximum input depth value pmax and the minimum input depth value pmin by the following equation.

  In Expression (1), qmax and qmin are the maximum output depth value and the minimum output depth value, respectively, and are predetermined constants. Here, the maximum output depth value and the minimum output depth value indicate the maximum value and the minimum value of the depth map output from the image processing apparatus, respectively. These values may be determined in advance according to a depth range that can be reproduced by the display device, or may be set individually by the user. Then, the depth conversion amount calculation unit 12 outputs the depth map, the input depth minimum value, the output depth minimum value, and the depth conversion amount to the depth conversion unit 13.

  The depth conversion unit 13 includes a depth structure map generation unit 131, a cut surface image generation unit 132, and a cutting processing unit 133. The depth structure map generation unit 131 generates a depth structure map Z (x, y, d) from the depth map P (x, y). (X, y) indicates the position coordinate of the pixel, and d indicates the depth of the pixel. If m and n are the number of horizontal pixels and the number of vertical pixels, respectively, the depth structure map Z (x, y, d) is 0 ≦ x ≦ m−1, 0 ≦ y ≦ n−1, and pmin ≦ d. It is a three-dimensional array defined in the range of ≦ pmax.

  The depth structure map Z (x, y, d) is calculated by the following equation, and represents the frequency with which a pixel having a depth of d exists in the vicinity of the pixel at the coordinate (x, y) and the vicinity thereof. However, weighting is performed not by the frequency itself but by g (i, j).

  w1, w2, and σ are predetermined positive constants. The depth structure map generation unit 131 outputs the obtained depth structure map Z (x, y, d) to the cut surface image generation unit 132.

  The cut surface image generation unit 132 generates a cut surface image that is an image representing the position of the cut surface. The cutting plane is a plane that cuts the depth structure map Z (x, y, d) along a plane perpendicular to the d-axis.

  FIG. 2A is a diagram illustrating an example of a cut surface. Reference numeral 21 denotes a range where a depth structure map exists, and reference numeral 22 denotes a cut surface. The cutting plane does not need to be a plane, and may pass through one d position for each coordinate (x, y) and different d positions depending on the coordinates (x, y). However, it is assumed that the condition that the difference absolute value of the value of d passing through the cut surface is equal to or less than the threshold value between the pixel at the coordinate (x, y) and the pixel adjacent thereto.

  Under this condition, the cutting plane is calculated so that the sum of the values of the depth structure map Z (x, y, d) at all (x, y) is minimized. That is, the cutting plane is calculated so that it passes through (x, y) and d where the frequency of pixels having the depth d exists in the vicinity thereof.

  Note that calculating the cut plane means obtaining a value of d through which the cut plane passes for each coordinate (x, y). An image holding the value of d through which the cut surface passes for each coordinate (x, y) is defined as a cut surface image A (x, y). The cut surface image is first generated for y = 0 and then sequentially from y = 1 to n-1. An evaluation value U (x, y, d) is obtained for each y, and a cut surface image is generated using dynamic programming.

  The process at y = 0 is performed as follows. First, the evaluation value U (x, 0, d) is calculated by the following equation from x = 0 to x = m−1 in order.

  At this time, the value of d 'given the minimum value in min is stored in the array B (x, y, d). α and β are positive constants. Next, d that gives the minimum value from U (m−1, 0, d) for x = m−1 is δ, and δ is a cut surface at coordinates (m−1, 0). That is, A (m-1, 0) = δ. Then, the cutting plane is calculated by A (x, 0) = B (x-1, 0, A (x-1, 0)) in order from x = m-2 to x = 0.

  The process at y> 0 is performed as follows. First, the evaluation value U (x, y, d) is calculated by the following equation from x = 0 to x = m−1 in order.

  C is a large constant that can be regarded as infinite. At this time, the value of d ′ that is the minimum in min is stored in the array B (x, y, d). Next, d that gives the minimum value from U (m−1, y, d) for x = m−1 is δ, and δ is a cut surface at coordinates (m−1, y). That is, A (m−1, y) = δ. Then, the cutting plane is calculated by A (x, y) = B (x-1, y, A (x-1, y)) in order from x = m-2 to x = 0.

  By performing this procedure in order for each y, the value of the cut surface image A (x, y) is calculated for all coordinates (x, y). Finally, with respect to the depth structure map Z (x, y, d), Z (x, y, d) = C is set for all (x, y, d) satisfying d = A (x, y). C is a large constant that can be regarded as infinite.

  In the cut surface image thus obtained, the absolute value of the difference between A (x, y) and A (x-1, y) is not more than α for each coordinate (x, y). This is an effect obtained by limiting the range in which min is taken to d + α ≦ d ′ ≦ d−α in the equations (3) and (4). For each coordinate (x, y), the absolute difference between A (x, y) and A (x, y-1) is less than or equal to α. This is an effect obtained by setting the evaluation value to C when | d−A (x, y−1) |> α in Expression (4). That is, the cut surface image obtained in this way has a property of having continuous values between adjacent pixels.

  The processing of the cut surface image generation unit 132 described above is for the case where the depth conversion amount k is 1. If k> 1, this process is repeated k times to calculate k cut surface images. Then, the obtained k cut surface images are output to the cutting processing unit 133.

  The cutting processing unit 133 calculates a converted depth map Q (x, y) from the depth map P (x, y). When k = 1, calculation is performed using the following equation.

FIG. 2B is a diagram for explaining the outline of Expression (5). (A) is a drawing when y = y ₁ and (b) is a drawing when y = y ₂ . This figure shows a two-dimensional diagram of the relationship between P (x, y) and A (x, y) and Q (x, y) when y = y ₁ and y = y ₂ , respectively. In the present embodiment, different processing is performed in two areas X (hatched area) and Y (unhatched area) where P (x, y) is divided by A (x, y). In the above equation (5), it is shown that P (x, y) existing in the region Y changes by d = 1 to become Q (x, y). In FIG. 2B, ● represents P (x, y), ▲ represents Q (x, y), x represents A (x, y), and x and y represent the coordinates of each pixel. Is shown. Thus, a continuous value is not taken and it becomes a discrete value.

  In the case of k> 1, first, the first cut surface image is input to A (x, y) of Expression (5) to calculate Q (x, y). Then, the obtained Q (x, y) is substituted into P (x, y) in equation (5), and the second cut surface image is substituted into A (x, y) in equation (5). Q (x, y) is calculated again. This process is repeated, and the calculation of Expression (5) is performed k times using k cut surface images. As a result, the maximum depth value is reduced by k.

  Finally, the cutting processing unit 133 adds qmin−pmin to all the coordinates (x, y) for the obtained Q (x, y). As a result, a post-conversion depth map Q (x, y) is generated in which the minimum value and the maximum value are qmin and qmax, respectively.

  By the processing using the above formula (5), the depth range of the depth map is reduced by changing the depth value of a pixel having a depth value larger than the pixel value of the same coordinate in the cut surface image in the depth map. Note that equation (6) may be used instead of equation (5).

  In the case of Expression (6), there is an effect of increasing the minimum value of depth by 1. In the depth map, the depth range of the depth map is reduced by changing the depth value of a pixel whose depth value is smaller than the pixel value of the same coordinate in the cut surface image. In the case of k> 1, after performing the process of Expression (6) k times, qmax−pmax is added to all the coordinates (x, y) for the obtained Q (x, y). As a result, a post-conversion depth map Q (x, y) is generated in which the minimum value and the maximum value are qmin and qmax, respectively.

  Finally, the depth conversion unit 13 outputs the generated post-conversion depth map Q (x, y) to the output unit 14. The output unit 14 outputs the converted depth map Q (x, y) from the image processing apparatus 10.

  In Expression (5) of the cutting processing unit 133, 1 is subtracted from P (x, y) to obtain P (x, y) -1, but the amount to be subtracted is not limited to 1 and is determined in advance. A predetermined amount r may be subtracted to obtain P (x, y) -r. Similarly, in equation (6), P (x, y) is incremented by 1 to obtain P (x, y) +1, but the amount to be added is not limited to 1, and a predetermined amount r is added. P (x, y) + r may be used. When r is used in this way, the number of times the process of the cut surface image generation unit 132 is repeated is k / r times, and k / r cut surface images are generated. In addition, the number of times the processing of the cutting processing unit 133 is repeated is k / r times.

  Next, the procedure of image processing performed by the image processing apparatus 10 according to the present embodiment will be described with reference to FIG. When receiving the input of the depth map by the function of the depth map input unit 11 (step S1), the image processing apparatus 10 calculates the depth conversion amount by the function of the depth conversion amount calculation unit 12 (step S2). Next, a depth structure map is generated by the function of the depth structure map generation unit 131 (step S3). Then, the number of cut surface images calculated by the depth conversion amount is calculated by the function of the cut surface image generation unit 132 (step S4). Thereafter, the cutting process is executed by the function of the cutting processing unit 133, and a post-conversion depth map is calculated (step S5). Finally, a post-conversion depth map is output by the function of the output unit 14 (step S6).

  As described above, according to the present embodiment, the depth range can be reduced even in the case of an image having a substantially uniform depth existence frequency. Since a cut surface image generated as a depth value with a low frequency of pixels having the depth value in the vicinity is used, the influence on the depth distribution in the subject is small, and the depth distribution in the subject can be maintained. Furthermore, since a cut surface image having continuous values between adjacent pixels is used, the influence on the local change in the depth map is small, and the depth distribution in the subject can be maintained. The image processing apparatus according to the present embodiment can be used for depth conversion for performing stereoscopic display in accordance with the characteristics of the display device, for example.

  In the present embodiment, in Expression (5) of the cutting processing unit 133, the pixel values of the depth map P (x, y) and the cut surface image A (x, y) are compared for each pixel to perform case classification. ing. The process of extracting pixels whose pixel values of the depth map are larger than the pixel values of the cut surface image according to the case division corresponds to region division that divides the depth map into two regions. Depth is an example of depth, and the depth map corresponds to a depth image. The part of the depth structure map generation unit 131, the cut surface image generation unit 132, and the cutting processing unit 133 until the case is divided by the expression (5) or (6) is the area dividing unit of the image processing apparatus according to this embodiment. Yes, the part where the depth is changed by a predetermined amount in the expression (5) or (6) of the cutting processing unit 133 corresponds to the depth changing means.

  Regarding the change of the depth (depth, etc.), the depth may be changed by a predetermined amount in at least one of the two areas as in the following equation (7).

  As a modification of the image processing apparatus according to the present embodiment, the depth map input unit 11 may further input a luminance image I (x, y) describing the luminance value of the subject in the same area as the depth map. . When the luminance image and the depth map have different image sizes, the luminance image size is converted in advance to be the same as the depth map size. The luminance image I (x, y) is output to the depth conversion unit 13 together with the depth map. When the depth structure map generation unit 131 generates the depth structure map Z (x, y, d), the following expression is used instead of the expression (2).

  f (x, y, d) and g (i, j) are the same as in equation (2), and h (x, y) in equation (8) represents the intensity of the edge of the luminance image. Using the depth structure map Z (x, y, d) obtained in this way, the same processing as that described above is performed by the cut surface image generation unit 132 and the cutting processing unit 133. By using the edge of the luminance image in this way, it is possible to further reduce the influence on the local change in the depth map in the region near the edge.

(Second Embodiment)
The second embodiment of the present invention relates to a luminance conversion process for converting a range of luminance values. FIG. 4 is a functional block diagram showing a configuration example of an image processing apparatus according to the second embodiment of the present invention. The image processing apparatus 40 according to the present embodiment inputs a luminance image, performs luminance conversion processing, and outputs a converted luminance image. As illustrated in FIG. 4, the image processing apparatus 40 includes a luminance image input unit 41, a luminance conversion amount calculation unit 42, a luminance conversion unit 43, and an output unit 44. In the image processing apparatus of the first embodiment, the input is replaced with a luminance image from the depth map, and the contents of the processing are the same.

  Hereinafter, details of each unit in the image processing apparatus 40 will be described. The luminance image input unit 41 inputs a luminance image and outputs it to the luminance conversion amount calculation unit 42.

  First, the luminance conversion amount calculating unit 42 calculates the minimum value and the maximum value of luminance from the luminance image, and sets the minimum value pmin and the maximum input luminance value pmax, respectively. Next, the luminance conversion amount k is calculated from the maximum input luminance value pmax and the minimum input luminance value pmin by the following equation.

  Here, qmax and qmin are an output luminance maximum value and an output luminance minimum value, respectively. The maximum output luminance value and the minimum output luminance value are predetermined constants. Then, the luminance conversion amount calculation unit 42 outputs the luminance image, the input luminance minimum value, the output luminance minimum value, and the luminance conversion amount to the luminance conversion unit 43.

  The luminance conversion unit 43 includes a luminance structure map generation unit 431, a cut surface image generation unit 432, and a cutting processing unit 433. The luminance structure map generation unit 431 generates a luminance structure map Z (x, y, d) from the luminance image P (x, y). (X, y) indicates the coordinates of the pixel, and d indicates the luminance value of the pixel. When m and n are the number of horizontal pixels and the number of vertical pixels, respectively, the luminance structure map Z (x, y, d) is 0 ≦ x ≦ m−1, 0 ≦ y ≦ n−1, and pmin ≦ d. It is a three-dimensional array defined in the range of ≦ pmax.

  The luminance structure map Z (x, y, d) is calculated by the equation (2), and represents the frequency of the pixel having the luminance d in the vicinity of the pixel at the coordinate (x, y) and the vicinity thereof. The luminance structure map generation unit 431 outputs the obtained luminance structure map Z (x, y, d) to the cut surface image generation unit 432.

  The cut surface image generation unit 432 generates a cut surface image that is an image representing the position of the cut surface. The cut plane is a plane that cuts the luminance structure map Z (x, y, d) along a plane perpendicular to the d axis. Similar to the cut surface of the first embodiment, the cut surface does not have to be a plane, and passes through one d position for each coordinate (x, y) and passes through a different d position depending on the coordinates (x, y). Good. However, it is assumed that the condition that the difference absolute value of the value of d passing through the cut surface is equal to or less than the threshold value between the pixel at the coordinate (x, y) and the pixel adjacent thereto.

  Under this condition, the cutting plane is calculated so that the sum of the values of the luminance structure map Z (x, y, d) at all (x, y) is minimized. That is, the cutting plane is calculated so that (x, y) and a pixel having a luminance value d in the vicinity thereof pass through d where the frequency is low.

  Next, k cut surface images A (x, y) are generated by the same processing as in the first embodiment. The cut surface image obtained in this way has the property of having continuous values between adjacent pixels, like the cut surface image of the first embodiment. Then, the obtained k cut surface images are output to the cutting processing unit 433.

  The cutting processing unit 433 calculates a converted luminance map Q (x, y) from the luminance map P (x, y). When the same method as that of the first embodiment is used and Expression (5) is used, the luminance of the luminance image is changed by changing the luminance of the pixel whose luminance is larger than the value of the pixel at the same coordinate in the cut surface image. The range is reduced. On the other hand, when Expression (6) is used, the luminance range of the luminance image is reduced by changing the luminance of the pixel whose luminance is lower than the pixel value of the same coordinate in the cut surface image in the luminance image.

  Finally, the luminance conversion unit 43 outputs the generated converted luminance image Q (x, y) to the output unit 44. The output unit 44 outputs the converted luminance image Q (x, y) from the image processing device 40.

  Next, the procedure of image processing performed by the image processing apparatus 40 according to the present embodiment will be described with reference to FIG. When receiving the input of the luminance image by the function of the luminance image input unit 41 (Step S1), the image processing apparatus 40 calculates the luminance conversion amount by the function of the luminance conversion amount calculation unit 42 (Step S2). Next, a luminance structure map is generated by the function of the luminance structure map generation unit 431 (step S3). Then, the cut surface image is calculated by the number of sheets determined by the luminance conversion amount by the function of the cut surface image generation unit 432 (step S4). Thereafter, a cutting process is executed by the function of the cutting processing unit 433, and a converted luminance image is calculated (step S5). Finally, the converted luminance image is output by the function of the output unit 44 (step S6).

  As described above, according to the present embodiment, it is possible to reduce the luminance range even in the case of an image having a substantially uniform luminance frequency. Since a cut surface image generated as a luminance value with a low frequency of pixels having the luminance value in the vicinity is used, the influence on the luminance distribution in the subject is small, and the luminance distribution in the subject can be maintained. Furthermore, since a cut surface image having a continuous value between adjacent pixels is used, the influence on the local change of the luminance image is small, and the luminance distribution in the subject can be maintained. This apparatus can be used, for example, for luminance conversion for displaying an image in accordance with the characteristics of the display apparatus.

  In the present embodiment, in Expression (5) of the cutting processing unit 433, the pixel values of the luminance image P (x, y) and the cut surface image A (x, y) are compared for each pixel to perform case classification. ing. This process of extracting a pixel whose pixel value of the luminance image is larger than the pixel value of the cut surface image by dividing into cases corresponds to area division that divides the luminance image into two areas. Note that the luminance is one of the depths, and the luminance image corresponds to the depth image.

  In addition, the area dividing unit of the image processing apparatus is a part of the luminance structure map generation unit 431, the cut surface image generation unit 432, and the cutting processing unit 433 until the case is divided by the expression (5) or (6). The part of the cutting processing unit 433 that changes the brightness by the predetermined amount corresponds to the depth changing means.

  Note that the content of the modification of the first embodiment can also be applied to the second embodiment.

  As described above, according to the present embodiment, it is possible to reduce the depth range while maintaining the depth distribution in the subject even in the case of an image having a substantially uniform depth frequency.

  In this specification, the depth value of each pixel constituting the depth map and the luminance value of each pixel constituting the luminance image have been described as examples. However, the present invention is not limited thereto, and each of the image Numerical values assigned to pixels can be targeted. It can also be used for adjustments such as enlargement in addition to depth reduction.

  Each component of the present invention can be arbitrarily selected, and an invention having a selected configuration is also included in the present invention.

  The above processing and control can be performed by software processing using a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), or FPGA (Field Programmable Gate) hardware.

  In the above-described embodiment, the configuration and the like illustrated in the accompanying drawings are not limited to these, and can be changed as appropriate within the scope of the effects of the present invention. In addition, various modifications can be made without departing from the scope of the object of the present invention.

  Each component of the present invention can be arbitrarily selected, and an invention having a selected configuration is also included in the present invention. In addition, a program for realizing the functions described in the present embodiment is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into a computer system and executed to execute processing of each unit. May be performed. The “computer system” here includes an OS and hardware such as peripheral devices.

  Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.

  The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the above-described functions, or may be a program that can realize the above-described functions in combination with a program already recorded in a computer system. At least a part of the functions may be realized by hardware such as an integrated circuit.

  The present invention includes the following disclosure.

(Appendix)
(1) An image processing apparatus that reduces a depth range of a depth image, and an area dividing unit that divides the depth image into two areas, and a depth that changes a depth by a predetermined amount in at least one of the two areas. An image processing apparatus comprising: a changing unit;

  According to the above, the depth of the depth image can be appropriately reduced.

(2) The area dividing unit converts the depth image to generate a cut surface image having a continuous value between adjacent pixels, and the pixel values of the depth image and the cut surface image And a cutting processing unit that divides the depth image into two regions by extracting pixels whose pixel value of the depth image is larger than the pixel value of the cut surface image. The image processing apparatus according to (1).

  As described above, the cut surface image can be appropriately generated. In addition, the regions divided by the cut surface image can be easily combined.

(3) Further, it has a depth structure map generation unit that generates a three-dimensional depth structure map from the depth image, and the cut surface image generation unit has a minimum sum of values of the depth structure map in all pixels. The image processing apparatus according to (2), wherein a depth value is calculated.

(4) The image processing apparatus according to any one of (1) to (3), wherein the depth represents a depth of a stereoscopic image.

(5) The image processing apparatus according to any one of (1) to (3), wherein the depth represents a luminance value of the image.

(6) An image processing method for reducing a depth range of a depth image, in which a region dividing unit divides the depth image into two regions, and a depth changing unit includes at least one of the two regions. And a step of changing the depth by a predetermined amount.

(7) A program for causing a computer to execute the image processing method according to (6).

  The present invention can be used in an image processing apparatus.

10: Image processing device 11: Depth map input unit 12: Depth conversion amount calculation unit 13: Depth conversion unit 14: Output unit 131: Depth structure map generation unit 132: Cut surface image generation unit 133: Cut processing unit 21: Depth structure Map range 22: cut surface 40: image processing device 41: luminance image input unit 42: luminance conversion amount calculation unit 43: luminance conversion unit 44: output unit 431: luminance structure map generation unit 432: cut surface image generation unit 433: Cutting processing section

Claims (5)

An image processing apparatus for reducing a depth range of a depth image,
An area dividing unit for dividing the depth image into two areas;
A depth changing unit that changes a depth by a predetermined amount in at least one of the two regions;
An image processing apparatus comprising: The region dividing unit corresponds to a pixel value of the depth image and the cut surface image, and a cut surface image generation unit that converts the depth image to generate a cut surface image having a continuous value between adjacent pixels. And a cutting processing unit that divides the depth image into two regions by comparing each pixel and extracting a pixel whose pixel value of the depth image is larger than the pixel value of the cutting plane image. Item 8. The image processing apparatus according to Item 1. And a depth structure map generating unit that generates a three-dimensional depth structure map from the depth image,
The image processing apparatus according to claim 2, wherein the cut surface image generation unit obtains a depth value that minimizes a sum of values of the depth structure map in all pixels.   The image processing apparatus according to claim 1, wherein the depth represents a depth of a stereoscopic image. The image processing apparatus according to claim 1, wherein the depth represents a luminance value of the image.

Images