The present disclosure 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, which can perform a more appropriate parallax control. In the related art, a technique has been developed in which a stereoscopic image is displayed using a display device. A sense of depth of a subject reproduced by the stereoscopic image is varied depending on a viewing condition when a user views the stereoscopic image, or a capturing condition of the stereoscopic image. For this reason, depending situations, a subject too protrudes or is too depressed, and thus a sense of depth which is reproduced is unnatural, which causes a user to feel exhausted. Therefore, there have been many techniques for appropriately controlling a parallax of a displayed stereoscopic image. For example, as such techniques, there has been a technique in which a stress value indicating a stress extent on a user when viewing a stereoscopic image is obtained based on a parallax of each pixel of the stereoscopic image, and the parallax of the stereoscopic image is controlled so as to minimize the stress value (for example, refer to Japanese Unexamined Patent Application Publication No. 2011-055022). However, there are cases where a completely appropriate parallax control may not be performed in the above-described technique. For example, in the technique for performing a parallax control for minimizing a stress value, there are cases where a comfortable sense of depth can be reproduced in the entire screen of the stereoscopic image; however, a depth position of a subject to which a user pays attention on the stereoscopic image may not be appropriate. As such, if the depth position of the main subject to which the user pays attention is inappropriate, the user experiences eye fatigue when viewing the stereoscopic image. It is desirable to more appropriately perform a parallax control of a stereoscopic image. According to an embodiment of the present disclosure, there is provided an image processing apparatus including an attention region estimation unit that estimates an attention region which is estimated as a user paying attention thereto on a stereoscopic image; a parallax detection unit that detects a parallax of the stereoscopic image and generates a parallax map indicating a parallax of each region of the stereoscopic image; a setting unit that sets conversion characteristics for correcting a parallax of the stereoscopic image based on the attention region and the parallax map; and a parallax conversion unit that corrects the parallax map based on the conversion characteristics. The image processing apparatus may further include an image synthesis unit that corrects a parallax of the stereoscopic image based on the corrected parallax map. The image processing apparatus may further include a maximum and minimum parallax detection unit that detects a maximum value and a minimum value of parallaxes indicated by the parallax map and detects a parallax of the attention region based on the parallax map and the attention region. In this case, the setting unit may set the conversion characteristics based on the maximum value, the minimum value, and the parallax of the attention region. The setting unit may set the conversion characteristics such that the parallax of the attention region is converted into a parallax with a predetermined size set in advance. The setting unit may set different conversion characteristics for the attention region on the stereoscopic image and regions other than the attention region on the stereoscopic image. The setting unit may set the conversion characteristics of the attention region on the stereoscopic image such that a parallax is linearly converted in a predetermined parallax section including the parallax of the attention region. The image processing apparatus may further include a smoothening unit that smoothens the attention region or the conversion characteristics. According to another embodiment of the present disclosure, there is provided an image processing method or a program including estimating an attention region which is estimated as a user paying attention thereto on a stereoscopic image; detecting a parallax of the stereoscopic image and generating a parallax map indicating a parallax of each region of the stereoscopic image; setting conversion characteristics for correcting a parallax of the stereoscopic image based on the attention region and the parallax map; and correcting the parallax map based on the conversion characteristics. In the embodiments of the present disclosure, an attention region which is estimated as a user paying attention thereto on a stereoscopic image is estimated; a parallax of the stereoscopic image is detected and a parallax map indicating a parallax of each region of the stereoscopic image is generated; conversion characteristics for correcting a parallax of the stereoscopic image are set based on the attention region and the parallax map; and the parallax map is corrected based on the conversion characteristics. According to still another embodiment of the present disclosure, there is provided an image processing apparatus including an attention region estimation unit that estimates an attention region which is estimated as a user paying attention thereto on a stereoscopic image; a parallax detection unit that detects a parallax of the stereoscopic image and generates a parallax map indicating a parallax of each region of the stereoscopic image; a histogram generation unit that generates a histogram of a parallax of the stereoscopic image indicated by the parallax map by weighting a parallax of the attention region; a shift amount calculation unit that calculates a shift amount indicating a correction amount of the parallax of the stereoscopic image based on the histogram; and an image processing unit that corrects the parallax of the stereoscopic image by shifting at least one of a left eye image or a right eye image forming the stereoscopic image based on the shift amount. According to still another embodiment of the present disclosure, there is provided an image processing method or a program including estimating an attention region which is estimated as a user paying attention thereto on a stereoscopic image; detecting a parallax of the stereoscopic image and generating a parallax map indicating a parallax of each region of the stereoscopic image; generating a histogram of a parallax of the stereoscopic image indicated by the parallax map by weighting a parallax of the attention region; calculating a shift amount indicating a correction amount of the parallax of the stereoscopic image based on the histogram; and correcting the parallax of the stereoscopic image by shifting at least one of a left eye image or a right eye image forming the stereoscopic image based on the shift amount. In the embodiments of the present disclosure, an attention region which is estimated as a user paying attention thereto on a stereoscopic image is estimated; a parallax of the stereoscopic image is detected and a parallax map indicating a parallax of each region of the stereoscopic image is generated; a histogram of a parallax of the stereoscopic image indicated by the parallax map is generated by weighting a parallax of the attention region; a shift amount indicating a correction amount of the parallax of the stereoscopic image is calculated based on the histogram; and the parallax of the stereoscopic image is corrected by shifting at least one of a left eye image or a right eye image forming the stereoscopic image based on the shift amount. According to still another embodiment of the present disclosure, there is provided an image processing apparatus including a scene recognition unit that performs scene recognition for a stereoscopic image; a parallax detection unit that detects a parallax of the stereoscopic image and generates a parallax map indicating a parallax of each region of the stereoscopic image; a setting unit that sets conversion characteristics for correcting a parallax of the stereoscopic image based on the parallax map; and a parallax conversion unit that corrects the parallax map based on the conversion characteristics and a result of the scene recognition. According to still another embodiment of the present disclosure, there is provided an image processing method or a program including performing scene recognition for a stereoscopic image; detecting a parallax of the stereoscopic image and generates a parallax map indicating a parallax of each region of the stereoscopic image; setting conversion characteristics for correcting a parallax of the stereoscopic image based on the parallax map; and correcting the parallax map based on the conversion characteristics and a result of the scene recognition. In the embodiments of the present disclosure, scene recognition for a stereoscopic image is performed; a parallax of the stereoscopic image is detected and a parallax map indicating a parallax of each region of the stereoscopic image is generated; conversion characteristics for correcting a parallax of the stereoscopic image is set based on the parallax map; and the parallax map is corrected based on the conversion characteristics and a result of the scene recognition. According to the embodiments of the present disclosure, it is possible to more appropriately perform a parallax control of a stereoscopic image. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram illustrating an outline of the present disclosure. Fig. 2 is a diagram illustrating an appropriate parallax range. Fig. 3 is a diagram illustrating a configuration example of the image processing apparatus according to an embodiment. Fig. 4 is a flowchart illustrating an image conversion process. Fig. 5 is a diagram illustrating an example of the conversion function. Fig. 6 is a diagram illustrating an example of the conversion function. Fig. 7 is a diagram illustrating an image conversion. Fig. 8 is a diagram illustrating a conversion function for each region. Fig. 9 is a diagram illustrating another configuration example of the image processing apparatus. Fig. 10 is a flowchart illustrating an image conversion process. Fig. 11 is a diagram illustrating still another configuration example of the image processing apparatus. Fig. 12 is a flowchart illustrating an image conversion process. Fig. 13 is a diagram illustrating each region of the parallax map. Fig. 14 is a diagram illustrating generation of the histogram. Fig. 15 is a diagram illustrating an example of the stress function. Fig. 16 is a diagram illustrating an example of the stress function. Fig. 17 is a diagram illustrating an effect of the parallax control considering an attention region. Fig. 18 is a diagram illustrating still another configuration example of the image processing apparatus. Fig. 19 is a flowchart illustrating an image conversion process. Fig. 20 is a diagram illustrating an example of the conversion function. Fig. 21 is a diagram illustrating still another configuration example of the image processing apparatus. Fig. 22 is a flowchart illustrating an image conversion process. Fig. 23 is a diagram illustrating smoothening of the attention region. Fig. 24 is a diagram illustrating smoothening of the conversion function. Fig. 25 is a diagram illustrating still another configuration example of the image processing apparatus. Fig. 26 is a diagram illustrating a configuration example of a parallax conversion unit. Fig. 27 is a flowchart illustrating an image conversion process. Fig. 28 is a diagram illustrating generation of the synthesis parallax map. Fig. 29 is a diagram illustrating a configuration example of the computer. DETAILED DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments of the present disclosure will be described with reference to the drawings Embodiments of the invention will now be described with reference to the accompanying drawings, throughout which like parts are referred to by like references, and in which: FIRST EMBODIMENT OUTLINE OF PRESENT DISCLOSURE First, an outline of the present disclosure will be described with reference to Fig. 1. In the present disclosure, for example, a parallax control of a stereoscopic image including a left eye image L and a right eye image R is performed as shown in the upper part of Fig. 1. When the left eye image L and the right eye image R are input, a parallax map DM11 indicating parallaxes of the left eye image L and the right eye image R is first generated. In addition, for example, if a histogram HT11 of a parallax of each pixel on the stereoscopic image is generated from the parallax map DM11, a parallax distribution of each subject in the stereoscopic image can be found. In addition, the transverse axis of the histogram HT11 indicates parallax, that is, a depth of the subject, and the longitudinal axis indicates a frequency of each parallax, that is, the number of pixels of the parallax map, having the parallax. Particularly, the right direction of the transverse axis of the histogram HT11 indicates a direction of the parallax where a subject is positioned on the front side when viewed from a viewing user, and the left direction of the transverse axis indicates a direction of the parallax where a subject is positioned on the depth side when viewed from the viewer. In addition, a subject having the parallax of "0" is positioned on a display surface which displays a stereoscopic image, and a subject having the parallax of a positive value is positioned further on the front side than the display surface when viewed from the user. Here, for example, a region which is estimated as the user paying attention thereto on a stereoscopic image is referred to as an attention region, and a region on the parallax map DM11, placed at the same position as the attention region is assumed as a region AR11. In addition, it is assumed that distributions of the respective pixels in the region AR11 correspond to a region AR'll part in the histogram HTll. In this case, the parallax of the attention region on the stereoscopic image has a positive number, and thus the subject in the attention region is viewed on the front side of the display surface when viewed from the user. That is to say, the subject seems protruding. Generally, if a parallax of a subject to which the user pays attention is small, and the subject is positioned around the display surface, inconsistency between convergence and focus adjustment does not occur, and thus the user hardly feels exhausted. Therefore, in the present disclosure, the parallax map is corrected, which is a corrected parallax map DM12, such that a subject of the attention region is positioned around the display surface, and a parallax of each subject on the stereoscopic image has a value in an appropriate parallax range giving a lesser burden to a user. In a parallax histogram HT12 obtained from the corrected parallax map DM12, parallax distributions of the respective pixels in the region AR12 placed at the same position as the attention region correspond to the AR' 12 part of the histogram HT12, and this it can be seen that the subject of the attention region is positioned around the display surface. If the corrected parallax map DM12 is obtained in this way, an image conversion is performed for the left eye image L and the right eye image R such that parallaxes of the left eye image L and the right eye image R become parallaxes shown in the corrected parallax map DM12, and thereby final left eye image L1 and right eye image R1 are generated. In the image conversion, one image of the left eye image L and the right eye image R may be not converted and only the other image thereof may be converted, or both the left eye image L and the right eye image R may be converted and then new left eye image L' and the right eye image R' are generated. In addition, the appropriate parallax range giving a lesser burden to a user when viewing a stereoscopic image is determined depending on a viewing distance of the stereoscopic image or a display device size. For example, as shown in Fig. 2, it is assumed that a user views a stereoscopic image at a position of a viewing distance Ls from a display surface DS11 which displays the stereoscopic image, and a binocular gap of the user is de. In addition, a distance between the user and a location of the subject on the stereoscopic image, that is, a position where a stereoscopic image of the subject is generated, is Ld, and a convergence angle in a case where the distance Ld to the position where a stereoscopic image is generated is the same as the viewing distance Ls is β. In addition, it is assumed that a convergence angle relative to a location of a subject having a minimum value of a parallax in an appropriate parallax range, that is, a location of the subject on the deepest side, is αmin, and an convergence angle relative to a location of a subject having a maximum value of a parallax in an appropriate parallax range is αmax In a case where a user views a stereoscopic image on the display surface DS11 in the viewing condition shown in Fig. 2, it is generally said that the user can view the stereoscopic image comfortably if the following Expression (1) is satisfied. In addition, in Expression (1), a denotes a convergence angle relative to a location of the substrate on the stereoscopic image. [Expression 1] In addition, a relationship between the viewing distance Ls and the convergence angle β is expressed by the following Expression (2) [Expression 2] Therefore, if the above Expression (2) is modified, the following Expression (3) can be obtained. [Expression 3] In addition, in a manner similar to Expression (2), a relationship between the distance Ld where a stereoscopic image is generated and the convergence angle a is expressed by the following Expression (4). [Expression 4] Here, since a range taken by the convergence angle a is amin^ccd'), a. pixel Gl' of an image of a phase p of a parallax between a phase 1 of the parallax of the right eye image R and a phase 0 of the parallax of the left eye image L is generated as a pixel of the left eye image L' after the image conversion. That is to say, a calculation of the following Expression (6) is performed, and a pixel value IM(x')of the pixel Gl' is calculated. [Expression 6] In addition, in Expression (6), IL(XL) and IR(XR) respectively indicate pixel values of the pixel Gl and the pixel Gr. In addition, the phase p of the generated parallax is p=d'/(d'+|d-d'|). In this way, the image synthesis unit 25 performs the calculation shown in Expression (6) for each pixel on the right eye image R, generates the new left eye image L' having the parallax d' with the right eye image R, and outputs a stereoscopic image formed by the right eye image R' and the left eye image L' as a stereoscopic image after the parallax is adjusted. In addition, although an example where the right eye image R is used as the right eye image R' without conversion and the left eye image L is converted into the left eye image L' has been described, the left eye image L may be used as the left eye image L' without conversion, and the right eye image R may be converted into the right eye image R' . Further, the right eye image R and the left eye image L may be respectively converted into the right eye image R' and the left eye image L'. In this case, the attention region estimation unit 21 detects an attention region from each of the right eye image R and the left eye image L, and the parallax detection unit 22 generates a parallax map having the right eye image R as a reference and a parallax map having the left eye image L as a reference. In addition, conversion characteristics are defined from an attention region on the right eye image R and a parallax map having the right eye image R as a reference, and the parallax map is converted into a corrected parallax map. In the same manner, conversion characteristics are defined from an attention region on the left eye image L and a parallax map having the left eye image L as a reference, and the parallax map is converted into a corrected parallax map. In addition, each pixel on the right eye image R and a corresponding pixel on the left eye image L corresponding to the pixel are used, a pixel value of a pixel of an image which is different from a parallax and a phase of the right eye image R by half of the parallax d' of the pixel of the corrected parallax map having the right eye image R as a reference is obtained, and thereby the left eye image L' is generated. In the same manner, each pixel on the left eye image L and a corresponding pixel on the right eye image R corresponding to the pixel are used, a pixel value of a pixel of an image which is different from a parallax and a phase of the left eye image L by half of the parallax d' of the pixel of the corrected parallax map having the left eye image L as a reference is obtained, and thereby right eye image R1 is generated. In this way, in the image processing apparatus 11, an attention region is positioned around the display surface displaying a stereoscopic image based on the parallax map and the attention region, and conversion characteristics are defined such that a parallax of each pixel of the stereoscopic image becomes a parallax in the allowable parallax range, thereby performing parallax adjustment. Thereby, a parallax of the stereoscopic image can be more appropriately controlled, and thus a user can view the stereoscopic image more comfortably. As a result, it is possible to reduce eye fatigue of a user. SECOND EMBODIMENT CONFIGURATION EXAMPLE OF IMAGE PROCESSING APPARATUS Although a case where all the pixels of the parallax map are converted by the same conversion function (conversion characteristics) has been described, a different conversion function may be used in a region at the same position as an attention region on the parallax map and in a region a position different from the attention region. In this case, for example, as shown in Fig. 8, a region AR41 located at the same position as an attention region on a parallax map DM21 is specified. In addition, as shown in the lower part of the figure, a conversion function for a parallax of each pixel in the region AR41 and a conversion function of a parallax of each pixel outside the region AR41 are defined. In the example shown in Fig. 8, a bent line F31 indicates a conversion function of a parallax of each pixel in the region AR41, and a bent line F32 indicates a conversion function of a parallax of each pixel outside the region AR41. In addition, in the graph of the conversion function, the transverse axis indicates a parallax di of each pixel on the parallax map, and the longitudinal axis indicates a corrected parallax d0. In the conversion function indicated by the bent line F31, the attention parallax dat is converted into 0, and in the section from the parallax dats including the attention parallax dat to the parallax date, the conversion function is a first-order function with a predetermined slope. In addition, the parallax dats, the parallax date, or a slope of the conversion function in the section between the parallax dats and the parallax date may be predefined, or may be set by a user. In addition, the parallax dats and the parallax date may be defined based on a parallax of each pixel in the region AR41. In this case, for example, a minimum value and a maximum value of parallaxes of pixels in the region AR41 respectively become the parallax dats and the parallax date. In addition, in the conversion function indicated by the bent line F31, the maximum parallax dmax and the minimum parallax d^n of the pixels on the parallax map are respectively converted into an allowable maximum parallax dmax' and an allowable minimum parallax dmin1/ and thus the conversion function is continuous in the overall sections. In this example as well, the section between the minimum parallax dmin and the parallax datsr and the section between the parallax date and the maximum parallax dmax show a first-order function. In contrast, in the conversion function indicated by the bent line F32, the maximum parallax dmax and the minimum parallax dmin are respectively converted into an allowable maximum parallax dmax' and an allowable minimum parallax dmin' • In addition, the attention parallax dat is converted into 0. That is to say, the conversion function indicated by the bent line F32 is the same as the conversion function indicated by the bent line Fll in Fig. 5. As such, if the parallax map is converted using the conversion functions having different conversion characteristics inside and outside the region AR41 of the parallax map DM21 located at the same position as the attention region, it is possible to more appropriately control a parallax of the attention region of the stereoscopic image. Particularly, characteristics of sections around the attention region in the conversion function inside the region AR41 are linear functions (first-order functions), and thereby it is possible to suppress distortion of depth around the attention region. As such, in a case where a parallax map is converted into a corrected parallax map with different conversion characteristics according to regions of the parallax map, an image processing apparatus is configured, for example, as shown in Fig. 9. In addition, in Fig. 9, parts corresponding to the case in Fig. 3 are given the same reference numerals, and description thereof will be appropriately omitted. An image processing apparatus 61 in Fig. 9 is the same as the image processing apparatus 11 in Fig. 3 in that the image processing apparatus 61 includes the attention region estimation unit 21 to the image synthesis unit 25. However, in the image processing apparatus 61, an estimation result of the attention region obtained by the attention region estimation unit 21 is supplied to the maximum and minimum parallax detection portion 31 of the parallax analysis unit 23 and the parallax conversion unit 24. In addition, the parallax conversion unit 24 performs conversion of a parallax map for each region based on the estimation result of the attention region and the conversion characteristics. DESCRIPTION OF IMAGE CONVERSION PROCESS Next, an image conversion process performed by the image processing apparatus 61 will be described with reference to the flowchart of Fig. 10. In addition, the processes in steps S41 to S44 are the same as those in the steps Sll to S14 of Fig. 4, and thus description thereof will be omitted. However, in step S41, an estimation result of the attention region is supplied to the maximum and minimum parallax detection portion 31 and the parallax conversion unit 24 from the attention region estimation unit 21. In step S45, the setting portion 32 sets conversion characteristics based on the maximum parallax, the minimum parallax, and the attention parallax supplied from the maximum and minimum parallax detection portion 31. Specifically, for example, as described with reference to Fig. 8, a conversion function for a region located at the same position as the attention region on the parallax map and a conversion function for a region located at a position different from the attention region on the parallax map, are set. The setting portion 32 supplies the conversion function set for each region to the parallax conversion unit 24. In step S46, the parallax conversion unit 24 converts the parallax map from the parallax detection unit 22 into a corrected parallax map based on the conversion functions (conversion characteristics) from the setting portion 32 and the estimation result from the attention region estimation unit 21. That is to say, the parallax conversion unit 24 converts a parallax of each pixel of the parallax map into a corrected parallax using the conversion function for a region which is located at the same position as the attention region on the parallax map. In addition, the parallax conversion unit 24 converts a parallax of each pixel of the parallax map into a corrected parallax using the conversion function for a region which is located at a position different from the attention region on the parallax map. In this way, the corrected parallax is obtained using the conversion function set for each region of the parallax map, and thereby a corrected parallax map is generated. The corrected parallax map is generated, and, thereafter, the image conversion process finishes through a process in step S47. This process is the same as the process in step S17 of Fig. 4, and thus description thereof will be omitted. The image processing apparatus 61 sets a conversion function for each region of the parallax map, particularly, conversion functions for a region which is the same as the attention region and the other region, and coverts the parallax map into a corrected parallax map. Thereby, it is possible to more appropriately control a parallax of the attention region. CONFIGURATION EXAMPLE OF IMAGE PROCESSING APPARATUS Although a case where a parallax of the stereoscopic image is adjusted based on the corrected parallax map has been described in the above description, a parallax may be adjusted by calculating a shift amount using a parallax of each pixel of a stereoscopic image, particularly, a parallax of an attention region, and shifting the left eye image L and the right eye image R. In this case, an image processing apparatus is configured, for example, as shown in Fig. 11. In addition, in Fig. 11, parts corresponding to the case in Fig. 3 are given the same reference numerals, and description thereof will be appropriately omitted. An image processing apparatus 91 in Fig. 11 includes an attention region estimation unit 21, a parallax detection unit 22, a histogram calculation unit 101, a stress value calculation unit 102, a shift amount calculation unit 103, and an image processing unit 104. The histogram calculation unit 101 generates a parallax histogram for each region on a stereoscopic image based on a parallax map from the parallax detection unit 22 and an attention region from the attention region estimation unit 21, and supplies the parallax histogram to the stress value calculation unit 102. When the histogram is generated, a parallax of each pixel of the attention region is weighted. In a case of adjusting a parallax by shifting the left eye image L and the right eye image R by a predetermined amount based on the histogram supplied from the histogram calculation unit 101, the stress value calculation unit 102 calculates a stress value indicating stress which a user feels when viewing the stereoscopic image where the parallax is adjusted, and supplies the stress value to the shift amount calculation unit 103. The shift amount calculation unit 103 calculates a shift amount which minimizes the stress value supplied from the stress value calculation unit 102, and supplies the shift amount to the image processing unit 104. The image processing unit 104 shits the supplied left eye image L and right eye image R based on the shift amount from the shift amount calculation unit 103, and generates and outputs a left eye image L' and a right eye image R'. DESCRIPTION OF IMAGE CONVERSION PROCESS Next, an image conversion process performed by the image processing apparatus 91 will be described with reference to the flowchart of Fig. 12. In step S71, the attention region estimation unit 21 estimates an attention region based on the left eye image L and the right eye image R which has been supplied, and supplies the estimation result to the histogram calculation unit 101. For example, the attention region is estimated through a process such as face recognition or visual attention. Specifically, for example, in the face detection, coordinates of the vertices of a rectangular region including a person's face on the left eye image L or the right eye image R, or a size of the rectangular region is detected. In addition, for example, in a case where the image processing apparatus 91 is an imaging apparatus capturing a stereoscopic image, information or the like obtained through an automatic focus process when capturing a stereoscopic image may be used. In other words, since a region of a subject where focus is adjusted on the stereoscopic image is specified in contrast type automatic focusing, the region of the subject where focus is adjusted is an attention region to which a user pays attention. As such, in a case where an attention region is estimated using information regarding the automatic focusing, it is possible to adjust a parallax of the stereoscopic image with respect to a subject to which a user capturing the stereoscopic image pays attention. In addition, in step S71, the same process as the process in step Sll of Fig. 4 is performed. In step S72, the parallax detection unit 22 detects a parallax based on the supplied left eye image L and right eye image R, and supplies a parallax map obtained as a result thereof to the histogram calculation unit 101. In addition, in step S72, the same process as the process in step S12 of Fig. 4 is performed. In step S73, the histogram calculation unit 101 generates a histogram indicating a parallax distribution for each region on the stereoscopic image based on the parallax map from the parallax detection unit 22 and the attention region from the attention region estimation unit 21, and supplies the histogram to the stress value calculation unit 102. For example, if a parallax map DM41 shown in Fig. 13 is supplied, the histogram calculation unit 101, the histogram calculation unit 101 generates a parallax histogram of pixels for each region of the central region DC of the parallax map DM41, the region DL around the left end, and the region DR around the right end. Here, the region DC is the same region as a central region on the stereoscopic image, and the region DL and the region DR are the same regions as regions of left and right image frame parts (left and right ends) of the stereoscopic image. In addition, hereinafter, the region DC, the region DL, and the region DR are respectively referred to as a central region DC, a left image frame region DL, and a right image frame region DR. For example, if a histogram of the central region DC is generated, the histogram calculation unit 101 sequentially selects pixels in the central region DC on the parallax map DM41 as an attention pixel. If the attention pixel is a pixel outside the attention region, the histogram calculation unit 101 adds 1 to a frequency value of the histogram bin of to which a parallax of the attention pixel belongs, and, if the attention pixel is a pixel in the attention region, adds a predetermined weight value W (here, 1