WO2012153661A1

WO2012153661A1 - Image correction device, image correction display device, image correction method, program, and recording medium

Info

Publication number: WO2012153661A1
Application number: PCT/JP2012/061447
Authority: WO
Inventors: 張　小▲忙▼; 上野　雅史; 宮田　英利
Original assignee: シャープ株式会社
Priority date: 2011-05-06
Filing date: 2012-04-27
Publication date: 2012-11-15

Abstract

One embodiment of the present invention pertains to an image correction device (1) comprising a target region extraction unit (10) that extracts, from input image data for the current frame, target region data including the same specific characteristics as target region data extracted from input image data for the immediately preceding frame, using a motion vector between the current frame and the immediately preceding frame and by browsing a target characteristics database (31).

Description

Image correction apparatus, image correction display apparatus, image correction method, program, and recording medium

The present invention relates to an image correction apparatus that corrects image data, an image correction display apparatus, and an image correction method.

In recent years, display devices that display images (moving images and still images) have been increased in screen size and image quality. Accordingly, a technique for improving the image quality of an image displayed on a display has been demanded.

By the way, in the conventional technique for improving the image quality, it is general that the entire image is corrected uniformly. As described above, when the correction is performed uniformly on the entire image, the image quality of a part of the image is improved, but the image quality of the other part of the image may be deteriorated. For example, in order to improve image quality, noise reduction that removes noise components included in image data is generally performed. If noise reduction is performed uniformly on the entire image, natural correction can be performed on blue sky clouds and human skin in the image data, but details disappear on bricks and lawns, etc. The problem that it will be done occurs. Therefore, the conventional technology that uniformly corrects the entire image cannot sufficiently display the potential of a display with high performance.

Therefore, various techniques have been proposed in which the entire image is not corrected uniformly but in consideration of the characteristics of each area of the image.

For example, Patent Document 1 listed below discloses a technique relating to an imaging device that corrects the hue of a color information signal corresponding to a human face area out of color information signals of an image based on a calculated hue correction value. Has been.

FIG. 13 shows the main components of the imaging device disclosed in Patent Document 1. FIG. 13 is a block diagram showing main components of the imaging device 90 disclosed in Patent Document 1. As shown in FIG. As illustrated in FIG. 13, the imaging device 90 includes a face area extraction unit 91 that extracts a color information signal corresponding to a person's face area, and a hue correction value that corrects the hue of the color information signal corresponding to the person's face area. A hue correction value calculation unit 92 to be calculated, and a hue correction unit 93 to correct the hue of the color information signal corresponding to the face area based on the hue correction value are provided.

As described above, if the technique disclosed in Patent Document 1 is used, a facial area in an image is specified, and color correction is performed on the specified facial area. It is possible to appropriately perform at least correction of the face area of a person, compared with the case where correction is made uniformly.

Further, in Patent Document 2 listed below, a color is applied to a skin color region representing a skin portion exposed by a person, which is detected based on a high brightness region indicating a high temperature portion of a subject extracted from an infrared image. A technique relating to a digital camera that performs correction is disclosed.

As described above, if the technique disclosed in Patent Document 2 is used, an area corresponding to the exposed skin portion of a person is specified, and color correction is performed on the specified area. As compared with the case where the entire image is corrected uniformly, at least correction of the exposed skin region of the person can be performed appropriately.

Japanese Patent Publication “Japanese Patent Laid-Open No. 2004-180114 (published on June 24, 2004)” Japanese Patent Publication “Japanese Patent Laid-Open No. 2004-207953” (published on July 22, 2004)

However, in the technique described in Patent Document 1, the area corresponding to the color information signal corresponding to the skin color in the image occupies a predetermined range with respect to the entire image, and the aspect ratio of the range However, an area having a predetermined ratio (for example, length: width≈1: 1) is determined as a face area. For this reason, it is difficult to extract only the face area by the method of discriminating the area that satisfies the above conditions. That is, in the technique described in Patent Document 1, an area other than an area that actually includes a face among areas corresponding to a color information signal corresponding to skin color may be determined as a face area. For this reason, an area other than the face area is set as a correction target, and an unnatural correction that should not be made in the area other than the face area is performed on the skin color information signal corresponding to the area other than the face area. There was a problem that there was something.

Further, in the technique described in Patent Document 2, since all the exposed skin parts of the person are detected and corrected, the person's face and arms cannot be individually detected, and each part is not detected. There was a problem that appropriate correction could not be performed.

Furthermore, when the technique of Patent Document 2 is applied to the technique of Patent Document 1, the area corresponding to the color information signal corresponding to the skin color in the image occupies a size of a predetermined range with respect to the entire image, In addition, if the aspect ratio of the range is a region having a predetermined ratio (for example, length: width≈1: 1), and the region is a high-luminance region indicating a portion having a high temperature, a very high probability It becomes possible to specify the face area.

Therefore, it is possible to avoid a situation in which an area other than the face area is targeted for correction and an unnatural correction that should not be performed on the area other than the face area is performed on the color information signal.

However, it is not possible to specify a detailed area such as whether the skin area other than the face area is a hand or a foot.

That is, even if Patent Documents 1 and 2 are combined, the face area can be specified, but other areas cannot be specified, and appropriate correction according to areas other than the face area is performed. The problem that cannot be done.

As described above, although the conventional technique enables the specification of the face area, other image areas cannot be specified for each characteristic, so appropriate correction according to the characteristics of each image area is performed. I can't.

In addition, in the conventional technology, when a moving image is included in an image, there is a possibility that the region to be corrected is moved for each frame, so it is difficult to track and specify the region to be corrected. A uniform correction must be made on the entire screen.

The present invention has been made to solve the above-described problems, and its main purpose is to display data indicating a target area to be corrected from image data for each characteristic even when a moving image is included in the image. An object of the present invention is to provide an image correction apparatus that can extract and perform correction suitable for data indicating the extracted target area.

In order to solve the above problem, an image correction apparatus according to an aspect of the present invention includes a motion vector detection unit that detects a motion vector between frames of input image data, and target region data stored in a storage unit. Target area extraction means for extracting target area data including the specific features from the input image data by referring to a target feature database including feature data indicating characteristics specific to the area extracted as the storage area, and the storage unit The image correction is performed on the target area data extracted by the target area extraction means in the input image data by referring to the correction content database in which the correction contents for the target area data are defined. Correction processing means, and the target region extraction means is detected by the motion vector detection means. Using the motion vector between the current frame and the previous frame, the target area data including the same specific features as the target area data extracted from the input image data of the previous frame is input from the input image data of the current frame. It is characterized by extracting.

According to the above configuration, when extracting the target area data from the input image data, feature data indicating characteristics peculiar to the area (correction target area) extracted as the target area data included in the target feature database Can be extracted as the target region data. In order to extract input image data applicable to the feature data as the target area data, for example, when a face area is extracted as the target area data, the same color as the partial area of the body or the face of the same color as the face area It is possible to prevent erroneous detection of the wall area of the camera.

Further, correction based on the correction content included in the correction content database can be performed on the extracted target area data. The content of the correction may be incompatible with the area other than the correction target area while being compatible with the characteristics of the correction target area. Therefore, since the data other than the target area data is not subjected to inadequate correction, the same correction is uniformly performed on the entire input image data, thereby preventing an unnatural output image.

Therefore, target area data to be corrected can be extracted from the input image data, and correction suitable for the extracted target area data can be performed.

Further, the target region extracting means uses the motion vector between the current frame and the immediately preceding frame detected by the motion vector detecting means, from the input image data of the current frame to the input image data of the immediately preceding frame. If the target area data extracted from the input image data of the immediately preceding frame is included in the input image data of the current frame by extracting the target area data containing the same unique features as the target area data extracted from The target area data can be reliably extracted. In other words, the target area extraction means tracks whether the target area data extracted from the input image data of the immediately preceding frame is included in the input image data of the current frame. Therefore, it is possible to reduce the leakage of the target area data.

For this reason, even if there is target area data that cannot be extracted by referring to the target feature database from the input image data of the current frame, the target area extraction unit inputs the target area data that cannot be extracted in the previous frame. When included in image data, using the motion vector, it is possible to extract target area data having the same characteristic as the target area data that cannot be extracted.

Therefore, by using the motion vector, it is possible to reduce omission of extraction and erroneous extraction of target area data.

Note that, as the feature data in the feature database, for example, when the relative distance relationship between eyes, nose, and mouth when a human face is viewed from the front is determined, the target region data that cannot be extracted is determined as the target data. List area data that does not apply to the relative distance between eyes, nose, and mouth, that is, area data of a person's profile, and area data of a person's face that is partially hidden by an object Can do.

The input image data includes moving image data, as well as still image data such as when a still image is displayed on a part of a moving image or a moving image is displayed on a part of a still image. A combination with moving image data is also included.

An image correction method for an image correction apparatus according to an aspect of the present invention is an image correction method for an image correction apparatus that corrects an input image in order to solve the above-described problem. The input image data is obtained by referring to a target feature database including a motion vector detection step for detecting a motion vector and feature data indicating features specific to a region extracted as target region data stored in the storage unit. The target area extraction step for extracting the target area data including the specific features from the above, and the input by referring to the correction content database in which the correction contents for the target area data stored in the storage unit are defined Image correction for the target area data extracted in the target area extraction step in the image data Performing the correction processing step, and in the target region extraction step, using the motion vector between the current frame detected in the motion vector detection step and the immediately preceding frame, from the input image data of the current frame, It is characterized in that target area data including the same characteristic as the target area data extracted from the input image data of the immediately preceding frame is extracted.

According to the configuration described above, the same effects as those of the image correction device described above can be obtained.

In order to solve the above-described problem, an image correction apparatus according to the present invention is specific to a correction target region stored in a storage unit and a motion vector detection unit that detects a motion vector between frames of input image data. By referring to a target feature database including feature data indicating features, target region extraction means for extracting target region data including specific features from input image data, and for the target region data stored in the storage unit Correction processing means for performing image correction on the target area data extracted by the target area extraction means in the input image data by referring to a correction content database in which correction contents are defined, and the target area The extracting means is between the current frame detected by the motion vector detecting means and the immediately preceding frame. That by using a motion vector, from the input image data of the current frame, it is characterized by extracting the object region data including the same unique features as target region data extracted from the input image data of the previous frame.

Thus, there is an effect that target area data to be corrected is extracted from the input image data, and correction suitable for the extracted target area data can be performed.

It is a block diagram which shows the structure of the image correction apparatus which concerns on one Embodiment of this invention. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on one Embodiment of this invention. It is a figure which shows an example of the object characteristic database in the image correction apparatus shown in FIG. It is a figure which shows an example of the correction content database in the image correction apparatus shown in FIG. 4 is a flowchart illustrating an example of a flow of correction target area extraction processing in a target area extraction unit of the image correction apparatus illustrated in FIG. 2. 3 is a flowchart illustrating an example of a flow of image correction processing in a correction processing unit of the image correction apparatus illustrated in FIG. 2. FIG. 4 is a diagram showing color difference correction for hue information of face area data in the CbCr coordinate system in the color difference correction processing unit of the image correction apparatus shown in FIG. 2, and (a) shows the range of skin color hue information of the face area data in CbCr coordinates; (B) is a diagram showing a range of hue information before and after correcting the skin color of the face area data in the CbCr coordinate system. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on the modification of one Embodiment of this invention. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on the other modification of one Embodiment of this invention. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on other embodiment of this invention. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on the modification of other embodiment of this invention. It is a block diagram which shows the detail of a structure of the image correction apparatus which concerns on the other modification of other embodiment of this invention. FIG. 10 is a block diagram illustrating main components of a face color correction device disclosed in Patent Document 1.

<Embodiment 1>
An image correction apparatus according to an embodiment of the present invention will be described with reference to FIGS. However, unless otherwise specified, the configuration described in this embodiment is not merely intended to limit the scope of the present invention, but is merely an illustrative example.

(Configuration of image correction device)
The configuration of the image correction apparatus according to the present embodiment will be described with reference to FIGS. 1 and 2. FIG. 1 is a block diagram illustrating a configuration of an image correction apparatus 1 according to the present embodiment. FIG. 2 is a block diagram showing details of the configuration of the image correction apparatus 1 according to the present embodiment.

The image correction device 1 is mounted on an image correction display device such as a television receiver or an information processing device that includes a display unit (not shown) that displays the input image data corrected by the image correction device. Has been.

That is, the image correction apparatus 1 is mounted on, for example, a television receiver or an information processing apparatus, and corrects the image quality of input image data included in a broadcast signal or an output signal of the image output apparatus. For this purpose, as shown in FIG. 1, the image correction apparatus 1 includes a target area extraction unit 10 (target area extraction unit), a correction processing unit 20 (correction processing unit), a storage unit 30, a motion vector detection unit (motion vector detection). Means) 50 and a frame memory 51. As shown in FIG. 1, the storage unit 30 stores a target feature database 31 and a correction content database 32.

Further, as shown in FIG. 2, the image correction apparatus 1 further includes an RGB conversion unit 40 (first color space signal conversion means). The correction processing unit 20 of the image correction apparatus 1 includes a luminance correction processing unit 21 (luminance correction processing unit), a color difference correction processing unit 22 (color difference correction processing unit), and a noise reduction processing unit 23 (noise reduction processing unit). It is comprised including.

The target area extraction unit 10 is means for extracting correction target area data (target area data) indicating a target to be corrected from input image data. Specifically, the target feature database 31 stored in the storage unit 30 is referred to, and feature information (a region extracted as correction target region data (correction target region) that is predetermined in the target feature database 31 is unique. Data having a value within the range value indicated by the feature data indicating the feature of (2) is extracted from the input image data as correction target region data. The target feature database 31 will be described later.

Further, the target area extraction unit 10 extracts the correction target area data from the input image data that is currently input based on the motion vector supplied from the motion vector detection unit 50 described later. Region data having the same characteristic as the correction target region data extracted from the input image data input immediately before the input image data is extracted as correction target region data. That is, the target area extraction unit 10 determines whether area data having the same characteristic as the correction target area data extracted from the input image data input immediately before is included in the currently input image data. Tracking is performed using a motion vector, and if it is included, correction target area data is extracted. A specific method for extracting correction target area data will be described later.

The correction processing unit 20 is means for performing image quality correction processing (hereinafter also referred to as target region correction processing) on the correction target region data. Specifically, the correction processing unit 20 stores, in the storage unit 30, the correction content most suitable for the target indicated by the correction target region data with respect to the correction target region data extracted by the target region extraction unit 10. The target area correction process is executed based on the determined content with reference to the correction content database 32 being determined.

Here, the target area correction processing is, for example, a luminance signal including luminance information (brightness value) indicating the degree of brightness of the correction target area data included in the correction target area data, and the color of the correction target area data. It means that at least one of the color difference signals quantitatively indicating the perceptual difference is adjusted. Further, the color difference signal includes hue information (hue value) indicating the hue of the correction target area data and the attribute of the color to be characterized, and saturation information (saturation value) indicating the degree of vividness of the correction target area data. include. The correction content database 32 will be described later.

The luminance correction processing unit 21 is provided in the correction processing unit 20 and executes a correction process on the luminance information included in the luminance signal of the correction target region data extracted by the target region extraction unit 10. Specifically, the luminance correction processing unit 21 corrects the luminance information in the correction target area data, thereby correcting a thick outline (for example, a thick edge such as a face outline), a thin and sharp outline. For example, fine outline emphasis correction for emphasizing (for example, thin and sharp edges such as eyelashes) and texture emphasis correction for emphasizing texture (for example, thin edges such as lawn and brick) are performed.

Note that the luminance correction processing unit 21 preferably includes a band pass filter and a high pass filter. The luminance correction processing unit 21 performs thick contour emphasis correction using a bandpass filter, performs fine contour emphasis correction using a bandpass filter having a higher pass frequency band than the bandpass filter used for thick contour emphasis correction, Texture enhancement correction may be performed using a high-pass filter. Of course, the configuration of the luminance correction processing unit 21 is not limited to this.

As a result, necessary correction can be performed for each data included in the correction target area data with respect to the correction target area data, so that the same correction is performed on the entire correction target area data, resulting in an unnatural output image. Can be prevented.

The chrominance correction processing unit 22 is provided in the correction processing unit 20 and is chromaticity indicating hue information and chromaticity information indicating hue information included in the chrominance signal of the correction target region data extracted by the target region extraction unit 10. Execute correction processing for information. Specifically, the color difference correction processing unit 22 performs a hue correction process on the hue information of the correction target area data, and performs a saturation correction process on the saturation information of the correction target area data.

In the hue correction process, the value of the hue information included in the correction target area data is corrected to a value within an appropriate range that is predetermined for the specific feature of the correction target area data. Specific features include, but are not limited to, for example, face, lawn, and blue sky. Examples of the hue correction process may include, but are not limited to, a skin color hue correction process that corrects the facial hue within an appropriate range, a blue sky hue correction process, and a lawn hue correction process. Is not to be done.

Also, the saturation correction process is performed by multiplying the color difference signal by a positive coefficient. When the coefficient is larger than 1, the color is corrected so as to be vivid, and when the coefficient is smaller than 1, the color is corrected so as to become light. When the coefficient is 1, saturation correction is not performed.

Note that the hue correction process is performed using the equation (1).

Where r is the saturation and θ is the hue. The coefficient k takes an arbitrary value within the range of 0 ≦ k ≦ 1, and the smaller the value of k, the stronger the effect of hue correction, and the strongest effect of hue correction when k = 0, All hues included in the data are corrected to θ1. When k = 1, no hue correction is performed.

The noise reduction processing unit 23 removes noise in the luminance signal and the color difference signal included in the correction target area data. Specifically, the noise reduction processing unit 23 removes noise (flickering, roughness, etc.) in the luminance signal and removes noise (excess color information, etc.) in the color difference signal, thereby removing the noise in the correction target area data. Remove. The noise reduction processing unit 23 preferably includes a low-pass filter or a median filter.

Here, the median filter arranges the density values in the respective pixels of the mask having a mask size of n × n (n is a natural number) (for example, 3 × 3 or 5 × 5) in ascending order, The noise is removed by setting the output density of the target pixel. The median filter has a larger noise removal effect as the mask size is larger.

In addition, the low-pass filter sets a coefficient for each pixel of the mask of the mask size n × n so that the sum of the coefficients becomes 1 and the coefficient of the target pixel becomes the maximum, and weighted averaging processing is performed. By removing the noise. When the coefficients are uniform, the noise removal effect is maximized. If the coefficient of the pixel of interest is 1, other coefficients are 0, and the noise removal effect is lost. As a low-pass filter coefficient, for a target region that is flat and has few edges, such as a face and a blue sky, a filter coefficient having a large noise removal effect may be used. For a target region with many edges, a filter coefficient with a small noise removal effect may be used.

Thereby, for example, when the noise reduction processing unit 23 is a low-pass filter, it is possible to perform correction by weighted average processing on the correction target region data, and when the noise reduction processing unit 23 is a median filter. Correction that removes fine noise can be performed.

The storage unit 30 includes a target feature database 31 that is referred to when the target region extraction unit 10 extracts correction target region data, and a correction that is referred to when the target region correction process is executed in the correction processing unit 20. A readable and writable memory in which the content database 32 is stored.

The storage unit 30 is a program for operating a computer as the image correction apparatus 1, and serves as a computer-readable recording medium that records a program that causes the computer to function as each unit of the image correction apparatus 1. Also bears.

The RGB conversion unit 40 converts the color space signal of the color system indicating the input image data corrected by the correction processing unit 20 into a color space signal of another color system and outputs it as output image data. For example, the RGB conversion unit 40 converts the color space signal (Y, Cb, Cr) of the input image data input in the color system represented by the YCbCr color space into the color space signal (R, G) of the RGB color system. , B).

In the present embodiment, the RGB conversion unit 40 will be described by taking as an example a configuration for converting a color space signal of the color system expressed in the YCbCr color space into a color space signal of the RGB color system. The present invention is not limited to this. For example, a configuration for converting to a color space signal of the CIE L * a * b * color system may be adopted.

The motion vector detection unit 50 is a means for detecting a motion vector between frames from input image data. The motion vector detection unit 50 obtains information such as a moving speed, a moving direction, and a moving distance from the difference between the input image data of the current frame and the input image data of the frame immediately before the current frame stored in the frame memory 51. A motion vector including is detected. In addition, the motion vector detection unit 50 supplies the detected motion vector to the target region extraction unit 10.

The frame memory 51 functions as a storage device that temporarily stores input image data in units of frames. The input image data stored at a certain time in the frame memory 51 is read out by the motion vector detecting unit 50 when the input image data stored immediately after the certain time is input. That is, when the input image data of the current frame is input, the input image data of the immediately previous frame stored in the frame memory 51 is read by the motion vector detection unit 50.

(Target feature database)
The target feature database 31 and the correction content database 32 stored in the storage unit 30 will be described with reference to FIGS. 3 and 4. FIG. 3 is a diagram showing an example of the target feature database 31 in the present embodiment. FIG. 4 is a diagram showing an example of the correction content database 32 in the present embodiment.

As illustrated in FIG. 3, the target feature database 31 extracts, for example, face feature information and a blue sky that are referred to when extracting a human face as feature information of a region extracted as correction target region data from input image data. This is a database in which a plurality of feature information, such as blue sky feature information referred to when performing grass extraction, and lawn feature information referred to when extracting grass, is set in advance.

As the facial feature information, for example, a range value value of hue information indicating the skin color of the face, a range value value of saturation information indicating the saturation of the face, a range value of value of luminance information indicating the brightness of the face, and The range value of the relative distance relationship of facial features such as eyes, nose and mouth is set.

Further, as the blue sky feature information, for example, the range value of the hue information, the range value of the saturation information value, the range value of the luminance information value, the range value of the high frequency component, and the range of the position on the screen Value is set. In addition, the range value of the high frequency component in the blue sky feature information includes, for example, a threshold value indicating that the ratio of the high frequency component included is below a certain level (for example, a high frequency component other than the high frequency component caused by noise) A small value that indicates that the position is not set), and the range value of the position on the screen is a value of a predetermined position that specifies that the position is higher than the predetermined position on the screen Should be set.

Further, as the lawn feature information, for example, the range value of the hue information value, the range value of the saturation information value, the range value of the luminance information value, the range value of the high frequency component, and the range of the position on the screen Value is set. The range value of the high frequency component in the lawn feature information is set with a threshold value indicating that the ratio of the high frequency component included is equal to or greater than a certain value, and the range value of the position on the screen includes the screen It is only necessary to set a value of a predetermined position that specifies that the position is lower than the predetermined position above.

Thus, it is possible to extract a plurality of types of correction target area data and perform image correction corresponding to each of the extracted types of correction target area data. That is, a plurality of correction target areas can be corrected, and more natural correction can be performed on the input image data.

(Correction content database)
The correction content included in the correction content database includes luminance correction content indicating correction content for luminance information, color difference correction content indicating correction content for color difference information, and noise reduction content indicating correction content for both luminance information and color difference information. Includes at least one of them.

This makes it possible to perform corrections according to the characteristics peculiar to each correction target area data. For example, for the correction target area data that should not be subjected to noise reduction, the noise reduction content is not included in the correction content, so that the noise reduction can be prevented.

Specifically, as shown in FIG. 4, the correction content database 32 is referred to when correcting the face correction content, which is referred to when correcting a person's face in the correction target region data, and when correcting the blue sky. Correction content information such as the blue sky correction content and the lawn correction content referred to when correcting the lawn is predetermined. As described above, in the correction content database 32, values within an appropriate range are determined in advance as correction content information for the unique features of each correction target area data.

For example, as face correction content, hue correction processing is performed on hue information indicating parts other than the eyes, nose, and mouth, the hue is corrected to an appropriate range of standard skin color (for example, compression), and hue information indicating the lip part is referred to Select the color that corrects the hue information indicating the lip from several predetermined colors, perform hue correction, perform edge enhancement on eyes, nose and mouth, and edge enhancement on eyelashes Etc. are set.

Also, the blue sky correction content is set such that the hue information indicating the blue sky is corrected to the appropriate range of the standard blue, and the noise reduction is performed on the hue information indicating the blue sky.

Also, the lawn correction contents are set such that the hue information indicating the lawn is corrected to a proper range of standard green, and the high frequency component is emphasized for the hue information indicating the lawn.

According to the above configuration, the correction for the luminance information, the correction for the hue information and the saturation information, and the correction for the luminance information, the hue information and the saturation information are respectively performed as necessary on the characteristics of the correction target region. And natural correction can be performed on the input image data.

Of the above-described brightness correction content, color difference correction content, and noise reduction content, correction processing that is unnecessary or incompatible due to the characteristics of the correction target region is not corrected by the corresponding processing unit, and the signal is passed through. You can do it.

In addition, since the target area extraction unit 10 and the correction processing unit 20 can perform processing based on the luminance information and the color difference information, the luminance signal and the color difference signal constituting the input image data are represented by the luminance signal and the color difference signal. It is possible to perform processing without converting to a color space signal of a color system different from the displayed color system.

In addition, since the RGB conversion unit 40 is provided, the luminance signal and the color difference signal including the luminance information and the color difference information processed by the correction processing unit 20 are different from the color system represented by the luminance signal and the color difference signal. After being converted to a color space signal, it can be output as output image data.

(Correction target area extraction process)
Next, the flow of the correction target area extraction process in the target area extraction unit 10 will be described with reference to FIG. FIG. 5 is a flowchart illustrating an example of the flow of the correction target area extraction process in the target area extraction unit 10 included in the image correction apparatus 1 according to the present embodiment.

In the present embodiment, as shown in FIG. 5, the target area extraction unit 10 acquires the face feature information, the blue sky feature information, and the lawn feature information shown in FIG. 3 from the target feature database 31 as the feature information. As an example, the feature information acquired by the target region extraction unit 10 in the present invention is not limited to this. Moreover, the target region extraction unit 10 may adopt a configuration in which which feature information is acquired can be set in advance by a setting unit (not shown).

As shown in FIG. 5, when the input image data is supplied, the target area extracting unit 10 acquires feature information defined in the target feature database 31 stored in the storage unit 30 (step S1).

When the feature information is acquired, the target region extraction unit 10 compares the luminance information, the hue information, and the saturation information included in the luminance signal and the color difference signal of the supplied input image data with the feature information (step S2).

The target area extraction unit 10 is data in which the area data indicated by the luminance information, hue information, and saturation information included in the luminance signal and the color difference signal of the input image data has a value within each range value of the face feature information. Is determined based on the comparison result (step S3).

The target area extraction unit 10 determines that the area data indicated by the luminance information, hue information, and saturation information in the input image data is data having a value within each range value of the face feature information (in step S3). YES) It is determined whether or not the eye, nose and mouth features can be extracted from the area data (step S4).

If it is determined that the eye, nose, and mouth feature information can be extracted (YES in step S4), the target region extraction unit 10 determines that the relative distance relationship between the eyes, nose, and mouth is within the range of the face feature information. It is determined whether it is a value (step S5).

When it is determined that the relative distance relationship between the eyes, the nose, and the mouth is a value within the range value of the facial feature information (YES in step S5), the target region extraction unit 10 determines the luminance information, hue information, and saturation information. Region data having a value within each range value of the face feature information from the input image data indicated by, and the region data in which the relative distance relationship between the eyes, nose and mouth is a value within the range value of the face feature information Are extracted as face area data to be corrected (step S6).

If the target area extraction unit 10 determines that the feature information of the eyes, nose, and mouth cannot be extracted (NO in step S4), the relative distance relationship between the eyes, nose, and mouth is not within the range value of the face feature information. If it is determined (NO in step S5), or whether the luminance information, hue information, and saturation information of the supplied input image data have been compared with all feature information after extracting face area data to be corrected Is determined (step S7).

If it is determined that the luminance information, hue information, and saturation information of the supplied input image data are not compared with all the feature information (NO in step S7), the target area extraction unit 10 again performs the processing from step S2. repeat.

The target area extraction unit 10 determines that the area data indicated by the luminance information, the hue information, and the saturation information is not data having a value within each range value of the face feature information in the input image data (NO in step S3). ) In the region, it is determined whether or not the region data indicated by the luminance information, hue information, and saturation information is data that is a value within each range value of the blue sky feature information (step S8).

The target area extraction unit 10 determines that the area data indicated by the luminance information, the hue information, and the saturation information in the input image data is data having a value within each range value of the blue sky feature information (in step S8). YES) In the area data, it is determined whether or not a high frequency component is included in the area data that is a value within each range value of the blue sky feature information (step S9). For example, the ratio value of the high frequency component included in the area data is compared with the threshold value set in the range value of the high frequency component in the blue sky feature information, and the ratio of the high frequency component included in the area data is equal to or less than the threshold value. In this case, it may be determined that the high frequency component is not included.

When it is determined that the high frequency component is not included (NO in step S9), the target region extraction unit 10 determines that the region data determined not to include the high frequency component is above the image represented by the input image data. It is determined whether or not the data has a value within a range value indicating that it is located at (step S10). For example, the target area extraction unit 10 determines whether or not a predetermined percentage of the area data determined to contain no high-frequency components is above the center of the image represented by the input image data. The determination in step S10 may be performed by determining the above.

When it is determined that the area data determined not to contain the high frequency component is area data having a value within a range value indicating that the area data is located above the image represented by the input image data (YES in step S10) In the target area extraction unit 10, the area data indicated by the luminance information, the hue information, and the saturation information is a value within each range value of the blue sky feature information, does not include a high frequency component, and the input image data is Area data having a value within a range value indicating that it is located above the image to be represented is extracted as blue sky area data to be corrected (step S11).

When it is determined that the high-frequency component is included (YES in step S9), the target region extraction unit 10 is not region data having a value within a range value indicating that it is located above the image represented by the input image data. (NO in step S10), or after extracting the blue sky region data to be corrected, whether the luminance information, hue information and saturation information of the supplied input image data are compared with all feature information It is determined again whether or not (step S7).

If it is determined that the luminance information, hue information, and saturation information included in the luminance signal and the color difference signal of the supplied input image data are not compared with all the feature information (NO in step S7), the target region extraction unit 10 Repeats the process from step S2 again.

The target area extraction unit 10 is an area in which the luminance information, the hue information, and the saturation information are determined not to be within the range values of the face feature information in the input image data (NO in step S3), and each of the blue sky feature information Whether or not the area data indicated by the luminance information, the hue information, and the saturation information is data that is a value within each range value of the lawn feature information in the area determined not to be within the range value (NO in step S8). Is determined (step S12).

The target region extraction unit 10 determines that the region data indicated by the luminance information, hue information, and saturation information in the input image data is data having a value within each range value of the lawn feature information (in step S12). YES) In the area data, it is determined whether or not a high frequency component is included in the area data within each range value of the lawn feature information (step S13). For example, the ratio value of the high frequency component included in the area data is compared with the threshold value set for the range value of the high frequency component in the lawn feature information, and the ratio of the high frequency component included in the area data is equal to or greater than the threshold value. If it is, it may be determined that a high frequency component is included.

When it is determined that a high frequency component is included (YES in step S13), the target region extraction unit 10 has a region data determined to include a high frequency component below the image represented by the input image data. It is determined whether or not the data has a value within a range value indicating that it is located at (step S14). For example, the target area extraction unit 10 determines whether or not a predetermined percentage of the area data determined to include a high frequency component is below the center of the image represented by the input image data. By determining the above, the determination in step S14 may be performed.

When it is determined that the area data determined to contain a high frequency component is area data having a value within a range value indicating that the area data is located below the image represented by the input image data (YES in step S14) The target region extraction unit 10 is a region data in which the region data indicated by the luminance information, the hue information, and the saturation information has a value within each range value of the lawn feature information, and includes high frequency components. And region data having a value within a range value indicating that the input image data is located below the image represented by the input image data is extracted as lawn region data to be corrected (step S15).

The target area extraction unit 10 includes, in the input image data, area data that is determined that the area data indicated by the luminance information, the hue information, and the saturation information has data within each range value of the lawn feature information. If not (YES in step S12), if it is determined that a high frequency component is not included (NO in step S13), a value within a range value indicating that the input image data is positioned below the image is displayed. If it is determined that the region data is not included (NO in step S14), or after extracting the lawn region data to be corrected, luminance information, hue information, and saturation information of the supplied input image data are all feature information. Is again determined (step S7).

If it is determined that the luminance information, hue information, and saturation information included in the luminance signal and the color difference signal of the supplied input image data are compared with all the feature information (NO in step S7), the target region extraction unit 10 Executes the tracking extraction process of the correction target area data (step S16). The tracking extraction processing of the correction target area data will be described later.

The target area extraction unit 10 ends the correction target area extraction process after executing the tracking extraction process of the correction target area data.

In the present embodiment, a configuration in which luminance information, hue information, and saturation information are first compared with face feature information, then compared with blue sky feature information, and further compared with lawn feature information will be described as an example. However, the order of comparison is not limited to this, and the order may be changed as appropriate.

Further, when feature information other than face feature information, blue sky feature information, and lawn feature information is defined in the target feature database, luminance information, hue information and saturation information of the input image data, face feature information, A step of comparing feature information other than blue sky feature information and lawn feature information may be added.

In addition, when a plurality of feature information is defined in the target feature database, a configuration in which it is possible to select in advance which one of the plurality of feature information is to be subjected to the correction target region extraction processing. It may be adopted. As a result, by comparing the luminance information, hue information, and saturation information with the preselected feature information among a plurality of feature information, only the region that matches the preselected feature region is extracted as the correction target region. can do.

(Trace extraction processing of correction target area data)
As described above, the target area extraction unit 10 tracks whether or not the correction target area data extracted from the input image data of the immediately preceding frame is included in the input image data of the current frame using the motion vector. If so, the correction target area data is extracted. Note that the input image data of the immediately preceding frame is also referred to as immediately preceding input image data.

Here, the tracking extraction process (step S16 shown in FIG. 5) of the correction target area data executed by the target area extraction unit 10 will be described.

When the tracking extraction process of the correction target area data is started, the target area extraction unit 10 is first extracted from the current input image data from the correction target area data extracted from the previous input image data using the motion vector. Tracking correction target area data that may be generated. The corrected image area data extracted from the immediately preceding input image data and the generated tracking correction target area data are also referred to as immediately preceding correction target area data.

Next, the target area extraction unit 10 refers to the correction target area data extracted in the correction target area extraction process (steps S1 to S15 shown in FIG. 5) that extracts correction target area data with reference to the target feature database 31, and Compare with the previous correction target area data. That is, the target area extraction unit 10 determines whether or not area data other than area data having the same characteristic as the correction target area data extracted in the correction target area extraction process exists in the immediately previous correction target area data. To do.

The target area extraction unit 10 does not include area data other than area data having the same unique characteristics as the correction target area data extracted from the input image data of the current frame by the correction target area extraction processing in the previous correction target area data. In this case, the tracking extraction process of the correction target area data is terminated.

The target area extraction unit 10 converts correction target area data other than area data having the same characteristic as the correction target area data extracted from the input image data of the current frame by the correction target area extraction processing into the immediately previous correction target area data. If it is included, the correction target area data other than the area data having the same characteristic as the correction target area data extracted from the input image data of the current frame by the correction target area extraction process is input to the input image data of the current frame. Are extracted as correction target area data candidates.

Next, the target area extraction unit 10 determines whether or not area data having the same characteristic as the correction target area data candidate exists in the input image data of the current frame using the motion vector. When it is determined that area data having the same unique characteristics as the correction target area data candidate exists in the input image data of the current frame, the target area extraction unit 10 extracts the area data as correction target area data.

As described above, the target area extraction unit 10 performs the tracking extraction process of the correction target area data.

In this way, for example, when the input image data is moving image data including a plurality of frames, the target region extraction unit 10 applies the input image data of the (n−1) th (where n is an integer greater than or equal to 2) frame. Whether or not the correction target area data included is included in the input image data of the nth frame can be tracked based on the motion vector. Then, the target area extracting unit 10, when the area data having the same characteristic as the target area data included in the n−1 input image data is included in the input image data of the nth frame, Data can be extracted from the input image data of the nth frame as correction target area data.

In the present embodiment, the target area extraction unit 10 performs the tracking target extraction process of the correction target area data after executing the correction target area extraction process of extracting the correction target area data with reference to the target feature database. However, the present invention is not limited to this. For example, the target region extraction unit 10 may execute a correction target region extraction process of extracting correction target region data with reference to the target feature database after executing the tracking extraction processing of the correction target region data.

In this case, in the tracking extraction process, the area data having the same characteristic as the immediately previous correction target area data is extracted from the current input image data, and then is not extracted in the tracking extraction process in the correction target area extraction process. The correction target area data is extracted from the area data. Accordingly, the area data to be extracted in the correction target area extraction process can be reduced, so that the addition of the target area extraction unit 10 related to the correction target area extraction process can be reduced.

(Target area correction processing)
Next, the flow of target area correction processing in the correction processing unit 20 will be described with reference to FIGS. 6 and 7. FIG. 6 is a flowchart illustrating an example of the flow of target area correction processing in the correction processing unit 20 of the image correction apparatus 1 according to the present embodiment. In the present embodiment, a target area correction process performed on the correction target face area data extracted by the target area extraction unit 10 will be described as an example.

As shown in FIG. 6, when the correction target area data is supplied, the correction processing unit 20 acquires correction content information defined in the correction content database 32 stored in the storage unit 30 (step S21). .

The correction processing unit 20 supplies information regarding correction for luminance information and correction target area data among the acquired correction content information to the luminance correction processing unit 21. In addition, the correction processing unit 20 supplies information regarding correction for hue information and saturation information and correction target area data among the acquired correction content information to the color difference correction processing unit 22. Further, the correction processing unit 20 supplies information regarding noise removal (noise reduction) and correction target area data among the acquired correction content information to the noise reduction processing unit 23.

The luminance correction processing unit 21 performs luminance correction processing on the luminance information included in the correction target region data extracted by the target region extraction unit 10 (step S22). For example, when the correction target area data is face area data, the luminance correction processing unit 21 refers to the face correction content information, performs thick outline emphasis correction that emphasizes thick outlines of eyes, nose, and mouth, and narrows the eyelashes. Performs fine contour emphasis correction that emphasizes steep contours. Note that the luminance correction processing unit 21 does not have to perform correction when the correction content for the luminance information is not defined in the correction content information. The luminance correction processing unit 21 supplies the input image data to the noise reduction processing unit 23 regardless of whether or not the luminance correction processing has been performed.

The color difference correction processing unit 22 performs a hue correction for correcting the hue information included in the correction target region data extracted by the target region extraction unit 10 (step S23). For example, when the correction target area data is face area data, the color difference correction processing unit 22 refers to the face correction content information and performs the correction calculation using the above-described equation (1), whereby the eyes, nose, and mouth. Correct the hue of the part other than to the appropriate range of the standard skin color.

In addition, when the correction target area data is face area data, the color difference correction processing unit 22 refers to the face correction content information, and refers to the lip portion with reference to the lip color of the original image. The hue correction is performed by selecting and applying the most natural lip color from the colors.

Note that the color difference correction processing unit 22 does not need to perform correction when the correction content for the hue information is not defined in the correction content information.

Here, with reference to FIG. 7, the correction calculation for the hue information of the face area data in the CbCr coordinate system will be described. FIG. 7 is a diagram showing color difference correction for the hue information of the face area data in the CbCr coordinate system. FIG. 7A is a diagram showing the range of skin color hue information of the face area data in the CbCr coordinate system, and FIG. 7B is a diagram before and after correcting the skin color of the face area data in the CbCr coordinate system. It is a figure which shows the range of the hue information after performing.

As shown in FIG. 7A, the hue θ shown in the hue information of the face area data is a value in the range from θ1−Δθ1 to θ1 + Δθ2 with θ1 as the center, and the saturation shown in the saturation information r is a value within the range from r1 to r2. Further, as shown in FIG. 7B, the intersection of θ1 + Δθ2 and r1 is a, the intersection of θ1−Δθ1 and r1 is b, the intersection of θ1−Δθ1 and r2 is c, and θ1 + Δθ2 and r2 When the intersection point with d is d, the value of the skin color hue information of the face area data is a value in the range abcd.

The color difference correction processing unit 22 corrects the value of the hue θ from θ1−Δθ1 ′ to θ1 + Δθ2 ′ centering on θ1, as shown in FIG. (However, Δθ1 ≧ Δθ1 ′ and Δθ2 ≧ Δθ2 ′). That is, as shown in FIG. 7B, the color difference correction processing unit 22 sets the intersection of θ1 + Δθ2 ′ and r1 as e, sets the intersection of θ1−Δθ1 ′ and r1 as f, and sets θ1−Δθ1 ′ and r2. When the intersection point between and the angle θ1 + Δθ2 ′ and r2 is h, the hue information of the value in the range abcd becomes the value in the range efgh by performing the correction calculation using the equation (1). So that it can be corrected.

Thereby, the hue value can be corrected within the range of the coordinate area where the input image data becomes a natural image in the CbCr coordinate system.

The color difference correction processing unit 22 performs saturation correction for correcting the saturation information included in the correction target region data extracted by the target region extraction unit 10 (step S24). For example, when the correction target area data is face area data, the color difference correction processing unit 22 performs correction with reference to the face correction content information. Note that the color difference correction processing unit 22 may not perform correction when the correction content for the saturation information is not defined.

The noise reduction processing unit 23 performs noise reduction that is correction for removing noise in luminance information and color difference information included in the correction target region data extracted by the target region extraction unit 10 (step S25). For example, when the correction target area data is face area data, the noise reduction processing unit 23 performs correction by referring to the face correction content information. In addition, the noise reduction process part 23 does not need to perform correction | amendment, when the correction content regarding noise reduction is not defined.

The correction processing unit 20 determines whether or not the target region correction processing has been performed on all the correction target region data extracted by the target region extraction unit 10 (step S26).

If it is determined that the target area correction process has not been performed on all the correction target area data extracted by the target area extraction unit 10 (NO in step S26), the correction processing unit 20 has not performed the target area correction process. The processing from step S22 to S26 is executed for the correction target area data. That is, the correction processing unit 20 repeats steps S22 to S26 until it is determined that the target region correction processing has been performed on all the correction target region data extracted by the target region extraction unit 10.

For example, when blue sky area data and lawn area data are extracted in addition to the face area data in the target area extraction unit 10, a step is performed so that the target area correction processing is performed on the blue sky area data and the lawn area data. The steps from S22 to S26 are repeated.

If it is determined that the target area correction process has been performed on all the correction target area data extracted by the target area extraction unit 10 (YES in step S26), the correction processing unit 20 ends the target area correction process.

According to the above-described configuration, when the correction processing unit 20 extracts the correction target area data from the input image data, the data corresponding to the feature data indicating the characteristic specific to the correction target area included in the target feature database 31 is extracted. It can be extracted as correction target area data. Accordingly, in order to extract the input image data applicable to the feature data as the correction target area data, for example, when extracting a face area as the correction target area data, a part of the body of the same color as the face area or a face It is possible to prevent erroneous detection of the same color wall area.

Further, the correction processing unit 20 can perform correction based on the correction content included in the correction content database 32 on the extracted correction target region data. The content of the correction may be incompatible with the area other than the correction target area while being compatible with the characteristics of the correction target area. Accordingly, inadequate correction is not performed on data other than the correction target area data, so that the same correction is uniformly performed on the entire input image data, and an unnatural output image can be prevented.

According to the above configuration, the image correction apparatus 1 according to the present embodiment can extract the correction target area data to be corrected from the input image data, and can perform correction suitable for the extracted correction target area data.

In addition, since a device such as a television receiver or an information processing device in which the image correction apparatus 1 according to this embodiment is mounted can display input image data corrected by the image correction apparatus 1, a more natural image can be displayed. The input image data corrected to become can be displayed.

Further, the target area extraction unit 10 has the same characteristic as the target area data extracted from the input image data of the frame immediately before the input image data of the current frame based on the motion vector detected by the motion vector detection unit 50. The target area data can be extracted from the input image data of the current frame. That is, the target area extraction unit 10 tracks whether the target area data extracted from the input image data of the immediately preceding frame is included in the input image data that is currently input. Data can be extracted.

Therefore, even when there is target area data that cannot be extracted by referring to the target feature database from the input image data of the current frame, the target area extraction unit 10 determines that the target area data that cannot be extracted is the input image of the previous frame. If it has been extracted from the data, the target area data that cannot be extracted can be extracted based on the motion vector.

Therefore, by using the motion vector, it is possible to reduce omission of extraction and erroneous extraction of the target area data.

As feature data in the feature database, for example, when the relative distance relationship between eyes and nose and mouth when a human face is viewed from the front is determined, target area data that cannot be extracted includes the determined eyes, Area data that does not apply to the relative distance between the nose and mouth, that is, area data of a person's profile, and area data of a person's face that has been partially hidden by an object crossing .

For example, when the area image data of the person's profile is included in the input image data of the current frame, the target area extraction unit 10 refers to the target feature database including the characteristic data of the front of the person's face. In some cases, the area data of a person's profile cannot be extracted as correction target area data. Even in such a case, the target area extraction unit 10 can extract the human side profile area data as the correction target area data based on the motion vector supplied from the motion vector detection unit 50.

[Modification 1]
A modification of this embodiment will be described with reference to FIG. For convenience of explanation, components having the same functions as those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted. In the present embodiment, differences from the first embodiment will be mainly described.

(Configuration of image correction device)
FIG. 8 is a block diagram showing details of the configuration of the image correction apparatus 1a according to this modification. As shown in FIG. 8, the image correction apparatus 1 a according to the first embodiment except that the image processing apparatus 1 a includes a correction processing unit 20 a instead of the correction processing unit 20 and an RGB conversion unit 40 a instead of the RGB conversion unit 40. The configuration is the same as that of the image correction apparatus 1.

The correction processing unit 20a is means for performing target area correction processing on the correction target area data. Specifically, for the correction target region data extracted by the target region extraction unit 10 based on the target feature database 31 and the motion vector, the content of correction most suitable for the target indicated by the correction target region data is stored in the storage unit 30. Determination is made with reference to the stored correction content database 32, and target area correction processing is executed based on the determined content.

Further, as shown in FIG. 8, the correction processing unit 20a includes a luminance correction processing unit 21a, a color difference correction processing unit 22a, and a noise reduction processing unit 23a.

The luminance correction processing unit 21a is provided in the correction processing unit 20a, and executes correction processing on luminance information included in the luminance signal of the correction target region data extracted by the target region extraction unit 10 from the input image data. In addition, the luminance correction processing unit 21a supplies the input image data that has been subjected to the correction process on the correction target region data to the RGB conversion unit 40a.

The color difference correction processing unit 22a is provided in the correction processing unit 20a, and executes correction processing on hue information and saturation information included in the color difference signal of the correction target region data extracted by the target region extraction unit 10 from the input image data. To do. Further, the color difference correction processing unit 22a supplies the input image data that has been subjected to the correction processing on the correction target area data to the RGB conversion unit 40a.

The RGB conversion unit 40a converts the color space signal of the color system indicating the input image data subjected to the correction process on the correction target area data in the luminance correction processing unit 21a and the color difference correction processing unit 22a into another color system. Convert to color space signal. For example, the RGB conversion unit 40 converts the color space signal (Y, Cb, Cr) of the input image data input in the color system represented by the YCbCr color space into the color space signal (R, G) of the RGB color system. , B). Further, the RGB conversion unit 40a supplies the input image data of the converted color space signal to the noise reduction processing unit 23a.

The noise reduction processing unit 23a is provided in the correction processing unit 20a, and removes noise in the luminance signal and the color difference signal included in the correction target region data from the input image data supplied from the RGB conversion unit 40a. The noise reduction processing unit 23a outputs the input image data from which noise has been removed as output image data.

(Correction target area extraction processing, target area correction processing)
The correction target area extraction process in the present modification is the same as the correction target area extraction process in the first embodiment shown in FIG.

In the target area correction process in this modification, the input image data is converted by the RGB conversion unit 40a between the saturation correction process (step S24) and the noise reduction process (step S25) shown in the target area correction process of FIG. This is the same as the target area correction processing according to the first embodiment, except that a step of converting the color space signal of the color system shown into a color space signal of another color system is included.

Thereby, the noise reduction processing unit 23a can perform the noise reduction process using the color space signal of the color system converted from the color system represented by the color space signal of the luminance signal and the color difference signal.

[Modification 2]
Another modification of the present embodiment will be described with reference to FIG. For convenience of explanation, components having the same functions as those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted. In the present embodiment, differences from the first embodiment will be mainly described.

(Configuration of image correction device)
FIG. 9 is a block diagram showing details of the configuration of the image correction apparatus 1b according to the present modification. As illustrated in FIG. 9, the image correction apparatus 1 b has the same configuration as the image correction apparatus 1 according to the first embodiment except that the image correction apparatus 1 b includes a correction processing unit 20 b instead of the correction processing unit 20.

The correction processing unit 20b is means for performing target area correction processing on the correction target area data. Specifically, for the correction target region data extracted by the target region extraction unit 10 based on the target feature database 31 and the motion vector, the content of correction most suitable for the target indicated by the correction target region data is stored in the storage unit 30. Determination is made with reference to the stored correction content database 32, and target area correction processing is executed based on the determined content.

As shown in FIG. 9, the correction processing unit 20b includes a luminance correction processing unit 21b, a color difference correction processing unit 22b, and a noise reduction processing unit 23b.

The noise reduction processing unit 23b is provided in the correction processing unit 20a, and removes noise in the luminance signal and the color difference signal of the correction target region data extracted by the target region extraction unit 10 from the input image data. The noise reduction processing unit 23b supplies the input image data from which noise has been removed to the luminance correction processing unit 21b and the color difference correction processing unit 22b.

The luminance correction processing unit 21b is provided in the correction processing unit 20b, and the luminance included in the luminance signal of the correction target region data extracted by the target region extraction unit 10 out of the input image data supplied from the noise reduction processing unit 23b. Execute correction processing for information. In addition, the luminance correction processing unit 21b supplies input image data obtained by performing correction processing on the correction target area data to the RGB conversion unit 40b.

The color difference correction processing unit 22b is provided in the correction processing unit 20b, and the hue included in the color difference signal of the correction target region data extracted by the target region extraction unit 10 out of the input image data supplied from the noise reduction processing unit 23b. Correction processing for information and saturation information is executed. Further, the color difference correction processing unit 22b supplies an input image signal obtained by performing correction processing on the correction target region data to the RGB conversion unit 40b.

The target area correction process in the present modification is the same as the target area correction process shown in FIG. 6, after performing the noise reduction process (step S25), the luminance correction process (step S22), the hue correction process (step S23), and the saturation. The target area correction process is the same as that of the first embodiment except that the correction process (step S24) is performed.

Thereby, the correction processing unit 20b can perform correction on the luminance signal and correction on the color difference signal after the noise reduction processing is performed in the noise reduction processing unit 23.

<Embodiment 2>
Another embodiment of the present invention will be described with reference to FIG. For convenience of explanation, components having the same functions as those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted. In the present embodiment, differences from the first embodiment will be mainly described.

(Configuration of image correction device)
FIG. 10 is a block diagram showing details of the configuration of the image correction apparatus 2 according to this modification. As shown in FIG. 10, the image correction apparatus 2 has the same configuration as the image correction apparatus 1 according to the first embodiment except that the image correction apparatus 2 further includes a YCbCr conversion unit 41 (second color space signal conversion unit).

The YCbCr conversion unit 41 converts the color space signal of the color system indicating the input image data into a color space signal of another color system. For example, the YCbCr conversion unit 41 converts an RGB color system color space signal representing input image data into a color system color space signal represented by a YCbCr color space. The YCbCr conversion unit 41 supplies the input image data obtained by converting the color space signal to the target area extraction unit 10, the correction processing unit 20, the motion vector detection unit 50, and the frame memory 51.

(Correction target area extraction processing, target area correction processing)
The correction target area extraction process and the target area correction process in the present embodiment are the same as the correction target area extraction process in the first embodiment shown in FIG. 5 and the target area correction process in the first embodiment shown in FIG. Description is omitted.

With the above configuration, when the input image data is a color space signal having a color system different from the color system represented by the YCbCr color space represented by, for example, a luminance signal and a color difference signal, the YCbCr color space is used. A color space signal of a color system different from the represented color system can be converted into a luminance signal and a color difference signal representing the YCbCr color space. As a result, the input image data can be corrected by correcting the luminance signal and the color difference signal regardless of the color space signal of the color system that the input image data is.

(Configuration of image correction device)
FIG. 11 is a block diagram showing details of the configuration of the image correction apparatus 2a according to this modification. As illustrated in FIG. 11, the image correction apparatus 2 a includes a correction processing unit 20 a instead of the correction processing unit 20, an RGB conversion unit 40 a instead of the RGB conversion unit 40, and a YCbCr conversion unit 41. Except for this, the configuration is the same as that of the image correction apparatus 1 according to the first embodiment.

The YCbCr conversion unit 41 converts the color space signal of the color system indicating the input image data into a color space signal of another color system. For example, the YCbCr conversion unit 41 converts an RGB color system color space signal representing input image data into a color system color space signal represented by a YCbCr color space. In addition, the YCbCr conversion unit 41 supplies the input image data obtained by converting the color space signal to the target area extraction unit 10, the correction processing unit 20 a, the motion vector detection unit 50, and the frame memory 51.

As shown in FIG. 11, the correction processing unit 20a includes a luminance correction processing unit 21a, a color difference correction processing unit 22a, and a noise reduction processing unit 23a.

The RGB conversion unit 40a converts the color space signal of the color system indicating the input image data subjected to the correction process on the correction target area data in the luminance correction processing unit 21a and the color difference correction processing unit 22a into another color system. Convert to color space signal. For example, the RGB converter 40 converts the color space signal (Y, Cb, Cr) of the input image data input in the color system expressed in the YCbCr color space into the color space signal (R, G in the RGB color system). , B). Further, the RGB conversion unit 40a supplies the input image data of the converted color space signal to the noise reduction processing unit 23a.

(Configuration of image correction device)
FIG. 12 is a block diagram showing details of the configuration of the image correction apparatus 2b according to this modification. As illustrated in FIG. 12, the image correction apparatus 2 b is the same as the image correction apparatus 1 according to the first embodiment except that the image correction apparatus 2 b includes a correction processing unit 20 b instead of the correction processing unit 20 and further includes a YCbCr conversion unit 41. It is a configuration.

The YCbCr conversion unit 41 converts the color space signal of the color system indicating the input image data into a color space signal of another color system. For example, the YCbCr conversion unit 41 converts an RGB color system color space signal representing input image data into a color system color space signal represented by a YCbCr color space. The YCbCr conversion unit 41 supplies the input image data obtained by converting the color space signal to the target area extraction unit 10, the correction processing unit 20 b, the motion vector detection unit 50, and the frame memory 51.

As shown in FIG. 12, the correction processing unit 20b includes a luminance correction processing unit 21b, a color difference correction processing unit 22b, and a noise reduction processing unit 23b.

(Program, storage medium)
Each block of the image correction apparatus 1 may be realized in hardware by a logic circuit formed on an integrated circuit (IC chip), or may be realized in software using a CPU (Central Processing Unit). Good.

In the latter case, the image correction apparatus 1 includes a CPU that executes instructions of a program that realizes each function, a ROM (Read Memory) that stores the program, a RAM (Random Access Memory) that expands the program, the program, A storage device (recording medium) such as a memory for storing various data is provided. An object of the present invention is a recording medium on which a program code (execution format program, intermediate code program, source program) of a control program of the image correction apparatus 1 which is software for realizing the functions described above is recorded so as to be readable by a computer. This can also be achieved by supplying the image correction apparatus 1 and reading and executing the program code recorded on the recording medium by the computer (or CPU or MPU).

Examples of the recording medium include tapes such as magnetic tapes and cassette tapes, magnetic disks such as floppy (registered trademark) disks / hard disks, and disks including optical disks such as CD-ROM / MO / MD / DVD / CD-R. IC cards (including memory cards) / optical cards, semiconductor memories such as mask ROM / EPROM / EEPROM / flash ROM, or PLD (Programmable logic device) or FPGA (Field Programmable Gate Array) Logic circuits can be used.

Further, the program code may be supplied to the image correction apparatus 1 via a communication network. The communication network is not particularly limited as long as it can transmit the program code. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network (Virtual Private Network), telephone line network, mobile communication network, satellite communication network, etc. can be used. The transmission medium constituting the communication network may be any medium that can transmit the program code, and is not limited to a specific configuration or type. For example, in the case of wired such as IEEE 1394, USB, power line carrier, cable TV line, telephone line, ADSL (Asymmetric Digital Subscriber Line) line, infrared such as IrDA or remote control, Bluetooth (registered trademark), IEEE 8021 wireless, HDR (High Data) Rate), NFC (Near Field Communication), DLNA (Digital Living Network Alliance), mobile phone network, satellite line, terrestrial digital network, and the like.

[Additional Notes]
As described above, the image correction apparatus according to an aspect of the present invention is extracted as target region data stored in the storage unit and the motion vector detection unit that detects the motion vector between frames of the input image data. A target area extraction unit that extracts target area data including the specific feature from the input image data by referring to a target feature database including characteristic data indicating a characteristic peculiar to the area, and is stored in the storage unit. Correction processing means for performing image correction on the target area data extracted by the target area extraction means of the input image data by referring to a correction content database in which correction contents for the target area data are defined. The target area extracting means includes a current frame detected by the motion vector detecting means. Using the motion vector between the previous frame and the target image data including the same specific features as the target image data extracted from the input image data of the previous frame is extracted from the input image data of the current frame. It is a feature.

According to the above configuration, when the target area data is extracted from the input image data, data corresponding to feature data indicating characteristics specific to the correction target area included in the target feature database is used as the target area data. Can be extracted. In order to extract input image data applicable to the feature data as the target area data, for example, when a face area is extracted as the target area data, the same color as the partial area of the body or the face of the same color as the face area It is possible to prevent erroneous detection of the wall area of the camera.

The target feature database in the image correction apparatus according to an aspect of the present invention includes a plurality of types of feature data including the feature data, and the correction content database includes a plurality of types of correction for each of the plurality of types of target region data. The target region extraction means includes a plurality of types of target region data corresponding to each of the plurality of types of feature data and the motion vector, and inputs the frame immediately before the input image data of the current frame. The target area data having the same characteristic as the target area data extracted from the image data is extracted, and the correction processing means performs image correction corresponding to each of the target area data extracted by the target area extraction means. Is preferred.

According to the above configuration, the correction processing means uses the target feature database and the motion vector, and each of the unique characteristics possessed by the plurality of types of target area data extracted by the target area extraction means. Can be corrected. In other words, the correction processing unit can appropriately correct a plurality of correction target regions having different unique features, and thus can perform more natural correction on the input image data.

In the image correction apparatus according to an aspect of the present invention, the target region data includes a luminance signal and a color difference signal, and the correction content includes a luminance correction content indicating a correction content for the luminance signal and a correction for the color difference signal. It is preferable to include at least one of a color difference correction content indicating the content and a noise reduction content indicating the correction content for both the luminance signal and the color difference signal.

According to the above configuration, it is possible to perform corrections according to features specific to each of the target area data. For example, with respect to target area data that should not be subjected to noise reduction, noise reduction can be prevented from being performed by including the noise reduction content in the correction content.

The correction processing means in the image correction apparatus according to one aspect of the present invention is based on the luminance correction processing means for correcting the luminance signal of the target area data based on the luminance correction content, and on the color difference correction content. The color difference correction processing means for correcting the color difference signal of the target area data, the luminance signal supplied from the luminance correction processing means based on the noise reduction content, and the color difference correction processing means supplied from the color difference correction means Noise reduction processing means for correcting both of the color difference signals, and the luminance signal and the color difference signal of the input image data supplied from the correction processing means are represented by the luminance signal and the color difference signal. It is preferable to further include first color space signal conversion means for converting to a color space signal of a color system different from the color system.

According to the above configuration, since the correction processing unit includes the luminance correction processing unit, the color difference correction processing unit, and the noise reduction processing unit, the correction for the luminance value indicated by the luminance signal, the color difference signal is The correction for the hue value and the saturation value shown, and the correction for the luminance value, the hue value and the saturation value can be respectively performed according to the characteristics of the correction target area, and the input image data Can be corrected naturally.

Of the above-described luminance correction content, color difference correction content, and noise reduction content, correction processing that is unnecessary or non-conforming due to the characteristics of the correction target region is not corrected by the corresponding processing means, and the signal is passed through. You can do it.

In addition, since the target area extraction unit and the correction processing unit can perform processing based on the luminance signal and the color difference signal, the luminance signal and the color difference signal constituting the input image data are converted into the luminance signal. Processing can be performed without conversion to a color space signal of a color system different from the color system represented by the signal and the color difference signal.

Here, the color system represented by the color space signal of the luminance signal and the color difference signal is, for example, a color system represented by a YCbCr color space. Further, different color system color space signals include, for example, CIEL * a * b * color system color space signals (L * signal, a * signal, b * signal) and RGB color system colors. Spatial signals (R (Red: red), G (Green: green), B (Blue: blue))) can be used.

Further, by providing the first color space signal conversion means, the luminance signal and the color difference signal processed by the correction processing means are different from the color system represented by the luminance signal and the color difference signal. It can be output as output image data after being converted into a color space signal of a color system (for example, an RGB signal of an RGB color system).

The correction processing means in the image correction apparatus according to one aspect of the present invention is based on the luminance correction processing means for correcting the luminance signal of the target area data based on the luminance correction content, and on the color difference correction content. Color difference correction processing means for correcting the color difference signal of the target area data; and noise reduction processing means for correcting the color space signal supplied from the first color space signal conversion means based on the noise reduction content. The first color space signal converting means includes the luminance signal and the color difference signal of the input image data supplied from the correction processing means, and a color system represented by the luminance signal and the color difference signal; It is preferable to convert to a color space signal of a different color system.

According to the above configuration, the noise reduction processing unit performs noise reduction processing using the color space signal of the color system converted from the color system represented by the color space signal of the luminance signal and the color difference signal. It can be carried out.

In the image correction apparatus according to an aspect of the present invention, the correction processing unit includes a noise reduction processing unit configured to correct the luminance signal and the chrominance signal based on the noise reduction content, and the luminance correction content. Luminance correction processing means for correcting the luminance signal supplied from the noise reduction processing means, and color difference correction processing for correcting the color difference signal supplied from the noise reduction processing means based on the color difference correction content. And the luminance signal and the color difference signal of the input image data supplied from the correction processing means are of a color system different from the color system represented by the luminance signal and the color difference signal. It is preferable to further comprise first color space signal conversion means for converting to a color space signal.

According to the above configuration, after the noise reduction processing is performed in the noise reduction processing unit, the luminance signal and the color difference signal can be corrected.

An image correction apparatus according to an aspect of the present invention uses a color space signal of input image data input by a color space signal of a color system different from the color system represented by the luminance signal and the color difference signal as the luminance signal. It is preferable to further include second color space signal conversion means for converting the signal and the color difference signal.

With the above configuration, when the input image data is a color space signal of a color system different from the color system represented by the YCbCr color space represented by the luminance signal and the color difference signal, for example, A color space signal of a color system different from the color system represented by the YCbCr color space can be converted into the luminance signal and the color difference signal representing the YCbCr color space. Thus, regardless of the color space signal of the color system in which the input image data is input, the input image data can be corrected by correcting the luminance signal and the color difference signal.

It is preferable that the brightness correction processing means in the image correction apparatus according to one aspect of the present invention is a band pass filter and a high pass filter.

According to the above configuration, for example, the bandpass filter corrects the data indicating the contour of the specific target included in the target region data, and the high pass filter includes the specific target included in the target region data. It is possible to correct the data indicating the texture. As a result, necessary correction can be performed on the target area data for each data included in the target area data, so that the same correction is performed on the entire target area data, resulting in an unnatural output image. Can be prevented.

It is preferable that the color difference correction processing unit in the image correction apparatus according to an aspect of the present invention performs a correction operation on the saturation and hue in the CbCr coordinate system.

According to the above configuration, the saturation and hue values can be corrected in the appropriate range of the coordinate area so that the input image data becomes a natural image in the CbCr coordinate system.

In the image correction apparatus according to one aspect of the present invention, the noise reduction processing unit is preferably a low-pass filter or a median filter.

According to the above configuration, for example, when the noise reduction processing means is a low-pass filter, the target area data can be corrected by weighted average processing, and the noise reduction processing means is a median filter. Correction that removes fine noise can be performed.

In the image correction apparatus according to an aspect of the present invention, the color difference signal includes hue information and saturation information, and the color difference correction processing unit includes the value of the hue information in the target area data. It is preferable that the hue information is corrected by setting the value within an appropriate range determined in advance according to the unique feature, and the saturation information is corrected by multiplying the color difference signal by a positive coefficient.

The image correction display device according to an aspect of the present invention preferably includes the image correction device and a display unit that displays the input image data corrected by the image correction device.

According to the above configuration, it is possible to display the input image data in which the input image data is corrected so as to become a more natural image in the image correction device.

An image correction method for an image correction apparatus according to an aspect of the present invention is an image correction method for an image correction apparatus that corrects an input image as described above, and detects a motion vector between frames of input image data. By referring to a target feature database including feature data indicating features specific to a region extracted as target region data, stored in the storage unit, and stored in the storage unit, from the input image data. By referring to a target area extracting step for extracting target area data including features and a correction content database in which correction contents for the target area data are stored, which is stored in the storage unit, Correction processing for performing image correction on the target area data extracted in the target area extraction step And using the motion vector between the current frame detected in the motion vector detection step and the previous frame in the target region extraction step, the previous frame is extracted from the input image data of the current frame. The target area data including the same characteristic features as the target area data extracted from the input image data is extracted.

A program for causing a computer to operate as the image correction apparatus according to the present invention, which causes the computer to function as each unit of the image correction apparatus, and a computer recording the program A readable recording medium is also included in the scope of the present invention.

In addition, this invention is not limited to this embodiment mentioned above, A various change is possible in the range shown to the claim.

The image correction apparatus according to the present invention can be suitably applied to a television receiver, a personal computer, a car navigation system, a mobile phone, a digital camera, a digital video camera, and the like.

1, 2 Image correction device 10 Target area extraction unit (target area extraction means)
20 Correction processing unit (correction processing means)
21 Luminance correction processing unit (luminance correction processing means)
22 Color difference correction processing section (color difference correction processing means)
23 Noise Reduction Processing Unit (Noise Reduction Processing Means)
30 Storage Unit 31 Target Feature Database 32 Correction Content Database 40 RGB Conversion Unit (First Color Space Signal Conversion Unit)
41 YCbCr converter (second color space signal converter)
50 Motion vector detection unit (motion vector detection means)
51 Frame Memory 90 Imaging Device 91 Face Area Extraction Unit 92 Hue Correction Value Calculation Unit 93 Hue Correction Unit

Claims

Motion vector detecting means for detecting a motion vector between frames of input image data;
By referring to a target feature database including feature data indicating features specific to a region extracted as target region data, stored in the storage unit, target region data including the specific features is input from the input image data. A target area extracting means for extracting;
An image of the target area data extracted by the target area extraction unit of the input image data by referring to a correction content database in which correction contents for the target area data are determined, stored in the storage unit. Correction processing means for performing correction,
The target area extracting unit extracts from the input image data of the immediately preceding frame from the input image data of the immediately preceding frame using the motion vector between the current frame and the immediately preceding frame detected by the motion vector detecting unit. An image correction apparatus characterized by extracting target area data including the same characteristic as the target area data.
The target feature database includes a plurality of types of feature data including the feature data,
The correction content database includes a plurality of types of correction content for each of the plurality of types of target area data,
The target area extraction unit extracts a plurality of types of target area data corresponding to each of the plurality of types of feature data and the motion vector from the input image data of the frame immediately before the input image data of the current frame. Target area data having the same characteristic as the target area data extracted,
The image correction apparatus according to claim 1, wherein the correction processing unit performs image correction corresponding to each of the target area data extracted by the target area extraction unit.
The target area data includes a luminance signal and a color difference signal,
The correction content is at least one of luminance correction content indicating the correction content for the luminance signal, color difference correction content indicating the correction content for the color difference signal, and noise reduction content indicating the correction content for both the luminance signal and the color difference signal. The image correction apparatus according to claim 1, further comprising:
The correction processing means includes
Brightness correction processing means for correcting the brightness signal of the target area data based on the brightness correction content;
Color difference correction processing means for correcting the color difference signal of the target area data based on the color difference correction content;
Noise reduction processing means for correcting both the luminance signal supplied from the luminance correction processing means and the color difference signal supplied from the color difference correction processing means based on the noise reduction content,
A first color signal that converts the luminance signal and the color difference signal of the input image data supplied from the correction processing means into a color space signal of a color system different from the color system represented by the luminance signal and the color difference signal. The image correction apparatus according to claim 3, further comprising color space signal conversion means.
The correction processing means includes
Brightness correction processing means for correcting the brightness signal of the target area data based on the brightness correction content;
Color difference correction processing means for correcting the color difference signal of the target area data based on the color difference correction content;
Noise reduction processing means for correcting the color space signal supplied from the first color space signal conversion means based on the content of the noise reduction,
The first color space signal conversion unit converts the luminance signal and the color difference signal of the input image data supplied from the correction processing unit to a color system different from the color system represented by the luminance signal and the color difference signal. The image correction apparatus according to claim 3, wherein the image correction apparatus is converted into a system color space signal.
The correction processing means includes
Noise reduction processing means for correcting the luminance signal and the color difference signal based on the noise reduction content;
Luminance correction processing means for correcting the luminance signal supplied from the noise reduction processing means based on the luminance correction content;
Color difference correction processing means for correcting the color difference signal supplied from the noise reduction processing means based on the content of the color difference correction;
A first color signal that converts the luminance signal and the color difference signal of the input image data supplied from the correction processing means into a color space signal of a color system different from the color system represented by the luminance signal and the color difference signal. The image correction apparatus according to claim 3, further comprising color space signal conversion means.
A second color space for converting a color space signal of input image data inputted by a color space signal of a color system different from the color system represented by the luminance signal and the color difference signal into the luminance signal and the color difference signal 7. The image correction apparatus according to claim 4, further comprising a signal conversion unit.
The image correction apparatus according to claim 4, wherein the brightness correction processing unit is a band-pass filter and a high-pass filter.
9. The image correction apparatus according to claim 4, wherein the color difference correction processing unit performs a correction operation on saturation and hue in a CbCr coordinate system.
The image correction apparatus according to claim 4, wherein the noise reduction processing unit is a low-pass filter or a median filter.
The color difference signal includes hue information and saturation information,
The color difference correction processing means includes:
Correcting the hue information by setting the value of the hue information to a value within an appropriate range that is predetermined for the specific characteristics of the target area data;
Correcting the saturation information by multiplying the color difference signal by a positive coefficient;
The image correction apparatus according to claim 4, wherein the image correction apparatus is an image correction apparatus.
The image correction apparatus according to any one of claims 1 to 11,
And a display unit for displaying the input image data corrected by the image correction apparatus.
An image correction method of an image correction apparatus for correcting an input image,
A motion vector detection step for detecting a motion vector between frames of the input image data;
By referring to a target feature database including feature data indicating features specific to a region extracted as target region data, stored in the storage unit, target region data including the specific features is input from the input image data. A target area extraction step to be extracted;
An image of the target area data extracted in the target area extraction step of the input image data by referring to a correction content database in which correction contents for the target area data are determined, stored in the storage unit. A correction processing step for performing correction,
In the target region extracting step, using the motion vector between the current frame and the immediately preceding frame detected in the motion vector detecting step, extraction is performed from the input image data of the immediately preceding frame from the input image data of the immediately preceding frame. An image correction method characterized by extracting target area data including the same characteristic as the target area data.
A program for operating a computer as the image correction apparatus according to any one of claims 1 to 11, wherein the computer functions as each unit of the image correction apparatus.
A computer-readable recording medium on which the program according to claim 14 is recorded.