JP2005141523A - Image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To estimate the scene attributes of an image to process the image in a way that suits the scene attributes. <P>SOLUTION: This image processing method which records or displays images according to image data includes detecting face areas using a face detection means, creating image processing condition parameters according to the number of face areas detected, and processing images according to the image processing condition parameters. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image processing method for recording or displaying an image based on image data from a photographic original or a digital camera.

  Conventionally, when an image is recorded on a recording material based on image data obtained by imaging a photographic film, various image processes have been performed to improve the quality of the recorded image. For example, the image quality correction is performed on the photographed image based on the photographing information input at the time of photographing (for example, Patent Document 1). The shooting information includes shooting distance information, continuous shooting information, flash emission information, shutter speed information, subject luminance information, and the like, and edge (contour) enhancement and contrast enhancement are performed as image quality correction.

In addition, there is a method that changes the image processing method depending on the size of the face image area of the image data and the presence or absence of the area, and adjusts the image processing of the face area in comparison with the background area according to the face size (for example, Patent Document 2). ).
Japanese Unexamined Patent Publication No. 7-159904 JP-A-9-233423

  In this way, various image processing is performed to improve the quality of the image to be recorded, and the image processing of the face area is adjusted in comparison with the background area according to the face size. Image processing according to portraits, snapshots, group photos, etc. was not possible.

  The present invention has been made in view of the above points, and an object thereof is to provide an image processing method capable of estimating a scene attribute of an image itself and performing image processing suitable for the scene attribute.

  In order to solve the above-described problems and achieve the object, the present invention is configured as follows.

The invention according to claim 1 is an image processing method for recording or displaying an image based on image data.
The face area is detected by the face detection means,
Generating image processing condition parameters according to the number of detected face regions;
An image processing method characterized in that image processing is performed based on the image processing condition parameter.

  According to a second aspect of the present invention, in the image processing method according to the first aspect, the degree of enhancement of a frequency component higher than a reference frequency for edge enhancement is relatively increased as the number of the face regions is larger. is there.

  According to a third aspect of the present invention, in the image processing method according to the first aspect, the degree of enhancement of frequency components lower than the reference frequency of edge enhancement is relatively increased as the number of face regions is smaller. is there.

  4. The image processing method according to claim 1, wherein the larger the number of face regions, the smaller the radius of the smoothing process or the non-application of the smoothing process. It is.

  The invention according to claim 5 is the image processing method according to claim 1, wherein as the number of the face regions is smaller, the processing radius of the smoothing process is increased or the restriction on the smoothing is loosened. is there.

  A sixth aspect of the present invention is the image processing method according to the first aspect, wherein the larger the number of the face regions, the more the image whose gradation contrast is shifted to the high contrast side is applied.

  The invention according to claim 7 is the image processing method according to claim 1, wherein the smaller the number of the face regions, the more the image is applied with the gradation contrast shifted to the soft side.

  The invention according to claim 8 is the image processing method according to claim 1, wherein the defect removal processing is weakened or the defect removal processing is not applied as the number of the face regions increases.

  The invention according to claim 9 is the image processing method according to claim 1, wherein the defect removal processing is increased as the number of the face regions is smaller.

  According to a tenth aspect of the present invention, a face area having a smaller size than a predetermined one of the face areas is excluded from the number of face areas for determining image processing conditions. The image processing method according to any one of the above.

  The invention according to claim 11 is characterized in that a face area having a low high-frequency component is excluded from the number of face areas for image processing condition determination. The image processing method according to item 1.

  According to a twelfth aspect of the present invention, an image processing condition is changed according to a combination of the number of the face areas and the size information of the face areas per one. The image processing method according to the item.

The invention according to claim 13 is an image processing method for recording or displaying an image based on image data.
The face area is detected by the face detection means,
Among the detected face areas, select the maximum size face area,
Generating an image processing condition parameter based on the size information of the selected face area;
An image processing method characterized in that image processing is performed based on the image processing condition parameter.

The invention according to claim 14 is an image processing method for recording or displaying an image based on image data.
The face area is detected by the face detection means,
Determine the sharpness of each detected face area;
Select the face area of the maximum size from those whose sharpness is a predetermined level or higher,
Generating an image processing condition parameter based on the size information of the selected face area;
An image processing method characterized in that image processing is performed based on the image processing condition parameter.

  With the above configuration, the present invention has the following effects.

  According to the first aspect of the present invention, image processing condition parameters are generated according to the number of detected face regions, and image processing is performed based on the image processing condition parameters. In this way, it is possible to estimate scene attributes of images such as non-person photographs, ID photographs, portraits, snapshots, group photographs, etc., and perform image processing suitable for the scene attributes.

  In the invention according to claim 2, as the number of face regions increases, the degree of enhancement of a frequency component higher than the reference frequency for edge enhancement is relatively increased, the scene attribute of the image itself such as a group photo is estimated, and the scene attribute It is possible to perform edge enhancement suitable for.

  According to the third aspect of the present invention, the degree of emphasis of frequency components lower than the reference frequency for edge emphasis is relatively increased as the number of face regions is smaller, and images such as non-person photographs, ID photographs, portraits, snapshots, etc. It is possible to estimate the scene attribute of itself and perform edge enhancement suitable for the scene attribute.

  In the invention according to claim 4, as the number of face regions increases, the processing radius of the smoothing process is reduced or the smoothing process is not applied, the scene attribute of the image itself such as a group photo is estimated, and the scene attribute It is possible to perform a smoothing process suitable for the above.

  In the invention described in claim 5, as the number of face regions is smaller, the processing radius of the smoothing process is increased or the restriction of smoothing is relaxed, and images such as non-person photos, ID photos, portraits, snapshots, etc. It is possible to estimate its own scene attribute and perform smoothing processing suitable for the scene attribute.

  In the invention described in claim 6, as the number of face regions increases, the whole image is applied with the gradation contrast shifted to the high contrast side, and the scene attribute of the image itself such as a group photo is estimated, which is suitable for the scene attribute. It is possible to perform a high gradation contrast.

  In the invention described in claim 7, as the number of face regions is smaller, the whole image is applied with the gradation contrast shifted to the soft side, and the image itself such as a non-person photograph, ID photograph, portrait, snapshot, etc. is applied. It is possible to estimate the scene attribute and soften the gradation contrast suitable for the scene attribute.

  In the invention described in claim 8, the defect removal processing is weakened as the number of face regions increases, or the defect removal processing is not applied, the scene attribute of the image itself such as a group photo is estimated, and the defect suitable for the scene attribute is obtained. A removal process can be performed.

  In the invention described in claim 9, the smaller the number of face regions, the larger the defect removal processing is, and the scene attributes of images such as non-person photos, ID photos, portraits, snapshots, etc. are estimated and suitable for the scene attributes. It is possible to perform defect removal processing.

  In the invention described in claim 10, a face area having a size smaller than a predetermined size of the face area is excluded from the number of face areas for determining the image processing condition, and the scene attribute of the image such as a group photo is estimated. It is possible to perform image processing suitable for the scene attribute.

  In the invention described in claim 11, face regions with low high-frequency components are excluded from the number of face regions for determining image processing conditions, and non-person photos, ID photos, portraits, snapshots, group photos It is possible to estimate the scene attribute of the image itself and to perform image processing suitable for the scene attribute.

  In the invention described in claim 12, the image processing conditions are changed by combining the number of face areas and the size information of the face area per one, and non-person photograph, ID photograph, portrait, snapshot, group photograph, etc. It is possible to estimate the scene attribute of the image itself and perform image processing suitable for the scene attribute.

  In the invention described in claim 13, the image processing condition parameter is generated based on the size information of the selected face area, the image processing is performed based on the image processing condition parameter, and the non-person photograph, ID photograph, portrait, snap It is possible to estimate scene attributes of images such as photographs and group photographs, and perform image processing suitable for the scene attributes.

  In the invention described in claim 14, the sharpness of each detected face area is determined, the face area of the maximum size is selected from those having a sharpness of a predetermined level or higher, and the image is based on the size information of the selected face area. Generate processing condition parameters, perform image processing based on the image processing condition parameters, estimate scene attributes of images such as non-person photos, ID photos, portraits, snapshots, group photos, etc., and images suitable for the scene attributes Processing can be performed.

  Hereinafter, although an embodiment of an image processing method of the present invention will be described, the present invention is not limited to this embodiment. The embodiment of the present invention shows the most preferable mode of the present invention, and the terminology of the present invention is not limited to this.

  FIG. 1 is a schematic configuration diagram of an image recording system for recording an image on a recording material based on image data, FIG. 2 is a diagram showing an original image, and FIG. 3 is a flowchart showing image processing. The image recording system 1 of this embodiment includes a scanner 3 that obtains image data from an original image 2, an image processing device 4 that performs image processing, and a printer 5 that performs image recording. In this embodiment, the printer 5 is provided, and an image is recorded on a recording material based on the image data. However, the image may be recorded on a storage medium / device or displayed on a display device.

  The means for obtaining the image data is not limited to the scanner 3 and may be image data from a CCD camera, a digital camera or the like. The image processing apparatus 4 is configured by a personal computer or the like, and performs image processing for recording an image on a recording material based on, for example, image data of an original image 2 of a snapshot as shown in FIG.

  The original image 2 is not limited to FIG. 2, but includes a non-person photograph, an ID photograph, a portrait, a group photograph, and the like. FIGS. 2A to 2D show original images 2 of four snap photographs, and show examples in which the number of persons is different.

  The image processing apparatus 4 performs processing as shown in FIG. In step s1, image data is obtained from the original image 2, and a face area is detected from the image data by face detection means (step s2).

  If there is no face area, color management processing is performed (step s3), image processing such as sharpness enhancement, noise removal, and defect removal is performed (step s4), and an output image is obtained (step s5).

  If there is a face area, an image processing condition parameter is generated (step s7) according to the number of detected face areas (step s6). Based on the image processing condition parameters, color management processing is performed (step s3), and image processing such as sharpness enhancement, contrast adjustment, noise removal, defect removal, edge enhancement, and smoothing processing is performed (step s4). An output image is obtained (step s5).

  For detecting this face area, color values such as saturation, hue value, and lightness are used alone or in combination, and a method in which an area in which a specific color is approximately continuous is used as a face area, or a technique such as Hough transform from an image is used. Known publicly known methods known in the art such as a method of extracting a specific shape pattern by extracting an edge to be used, a structure extraction method by template matching, or a combination of these can be employed.

Specifically, Japanese Patent Application Laid-Open No. 52-156624 discloses that a measurement point included in a predetermined color region is defined as a skin color on two-dimensional or three-dimensional color coordinates, and the number of measurement points determined to be a skin color. A method for determining that there is a flesh color in an image when there are 13 or more, Japanese Patent Application Laid-Open No. 4-346332 divides the original image into a number of pixels, and calculates the hue value and saturation value from the BGR value of each pixel. A method of creating a histogram, dividing the histogram from its shape, dividing an image into regions composed of pixels corresponding to the divided parts, and estimating a region corresponding to a face from the plurality of regions on the left, -309433, JP-A-6-67320, JP-A-5-158164, and JP-A-5-165120 determine skin color regions based on histograms such as hue value, saturation value, and luminance value, This skin color A method of making a region a face region, Japanese Patent Application Laid-Open No. 9-138471, a method of extracting contours by comparing edge strengths within a search range, measuring the density or luminance of a plurality of points in an image, and determining the amount of change, After setting a point where the amount of change is equal to or greater than a predetermined value as a reference point, a search range and a search direction pattern are set using a change amount such as density within a predetermined range from the reference point, and the search direction pattern is set within the search range. A method is illustrated in which a location where the amount of change in concentration or the like in the indicated direction is greater than or equal to a predetermined value is searched, and then the search is repeated using this location as a reference, and a specific location is extracted by connecting the searched and set reference points. The

  In addition, a method for extracting a skin color area and extracting a non-edge part area as a face area in JP-A-9-101579, and extracting a partial part in JP-A-11-316845 and combining them into main subject candidates A region having a high probability of being a region corresponding to a face based on a result obtained by weighting the results of a plurality of extraction methods in which weighting factors are individually set in Japanese Patent Application Laid-Open No. 9-138470. A method for extracting as a face area, Japanese Patent Laid-Open No. 8-63597, prepares a plurality of face templates, performs template matching between the face template and an image, extracts face area candidates, and then determines the skin color in the face area candidates. A method of extracting a face region based on a distribution, and dividing the image into a plurality of regions by subjecting an image to binarization or the like in JP-A-8-184925, and dividing from the plurality of regions , A method for extracting a region having the highest probability, a method disclosed in Japanese Patent Laid-Open Nos. 5-210739, 5-274438, 5-274439, 5-307537, and Japanese Laid-Open Patent Publication Nos. 5-307605, 5-282457, 6-214970, 6-309457, 7-234856, and 8-87589 are arranged by a neural network. The line process method is disclosed in Japanese Patent Application Laid-Open Nos. 8-221567, 2000-20694, 2000-32272, 2000-201358, and 2000-20769. Although there are methods such as the method used, this is an example and the present invention is not limited to this.

  In addition to the above method, the face area may be extracted by manually specifying the face area on the monitor by using a mouse, a light pen, or the like. You may use together with the extraction method mentioned.

  Also, the color management process is to avoid color images appearing differently for each device by expressing a color image using a common color space in a plurality of devices. That is, based on the basic concept that two colors represented by the same coordinate value in a certain color space look the same color to the human eye, all colors in the image are expressed in one color space, thereby It tries to match the appearance of the color of the image. For color management, a method of correcting differences in color appearance for each device by using XYZ three threshold values as the color space is proposed in Japanese Patent Laid-Open No. 2003-250055.

  The image processing based on the image processing condition parameters includes sharpness enhancement, contrast adjustment, noise removal, defect removal, edge enhancement, smoothing processing, and the like.

  The sharpness enhancement process is a process of sharpening an image as a method for removing “blur” which is one cause of image quality deterioration and making the image clear. The image blur is caused by the high spatial frequency component being weakened compared to the low spatial frequency component. The effect appears at the boundary portion (edge) of the uniform density region. Therefore, in order to remove the blur, processing for enhancing high spatial frequency components is performed, and a high frequency enhancement filter in the image space is used.

  As the sharpness enhancement processing, for example, as shown in FIG. 4, a convolution operation using a 3 × 3 or 5 × 5 size spatial filter can be performed.

  FIG. 4A shows an example of a mask for emphasizing particularly high frequency components of 3 × 3 size, and the degree of emphasis is adjusted by the value of k. FIG. 4B shows an example of a mask that particularly emphasizes low frequency components, and the degree of enhancement is adjusted by the value of k.

  In contrast adjustment processing, for example, contrast can be increased by density gradation conversion. Further, as a contrast enhancement method, there is a method in which the density value of an image corresponds to a color. For example, a lower value corresponds to red, a higher value corresponds to orange, and a higher value corresponds to yellow. There is a technique described in “Digital Image Processing” published by Modern Science Co., Ltd., and the like. Further, there is a method such as Japanese Patent Laid-Open No. 8-96141 that determines whether the contrast difference between two colors of a plane is the largest and controls to perform two-color shaping on the plane, but this is an example. However, the present invention is not limited to this.

  The noise removal process uses an ε-filter and determines whether or not the pixel is an edge part based on difference information from adjacent pixel data. If it is determined to be an edge portion, the original image data is used as it is. When it is determined as a non-edge portion, an average value with adjacent pixels is newly adopted as image data. The noise removal intensity can be adjusted by adjusting the conditions for determination and adjusting the radius of the average target area.

  In the defect removal processing, it is determined whether or not the pixel is any defect (image failure due to adhesion of scratches, dust, fingerprints, etc.) based on the response information at the time of infrared irradiation, and the image information is corrected according to the degree of the defect.

  For the edge enhancement processing, there is a technique such as Japanese Patent Laid-Open No. 5-307603 which slightly expands or sharpens the edge portion of the image data to the outside of the image for the pixel determined as the edge region. Is an example and the present invention is not limited to this.

  The smoothing process preferably uses a method of smoothing without causing blur. Further, when the position where the noise exists in the image is known, or when the noise in the image can be distinguished from other portions, the noise is removed by replacing the density value of the noise portion with the average density in the vicinity. In addition, there is a method of weakening noise by an averaging operation without distinguishing between noise and other parts in an image. Further, there is a technique such as Japanese Patent Laid-Open No. 8-18777 which performs smoothing processing using a filter having a weak enhancement and a flat characteristic in a specific frequency region, but this is an example, and the present invention is not limited to this. Absent.

  In this embodiment, a face area is detected by a face detection unit, an image processing condition parameter is generated according to the number of detected face areas, and image processing is performed based on the image processing condition parameter. It is possible to estimate scene attributes of images such as ID photos, portraits, snapshots, and group photos, and perform image processing suitable for the scene attributes.

  Image processing based on this image processing condition parameter is performed as follows.

  As for edge enhancement, for example, as shown in FIGS. 2C and 2D, the degree of enhancement of frequency components higher than the reference frequency for edge enhancement is relatively increased as the number of face regions increases. It is possible to estimate a scene attribute of an image itself such as a group photo from the number of face areas and perform edge enhancement suitable for the scene attribute.

  For example, as shown in FIGS. 2A and 2B, the degree of enhancement of frequency components lower than the reference frequency for edge enhancement is relatively increased as the number of face regions is smaller. It is possible to estimate the scene attribute of the image itself such as a non-person photograph, ID photo, portrait, snapshot, etc. from the number of face areas, and perform edge enhancement suitable for the scene attribute.

  As for the smoothing process, as shown in FIGS. 2C and 2D, for example, the larger the number of face regions, the smaller the process radius of the smoothing process or the non-application of the smoothing process. It is possible to estimate a scene attribute of an image itself such as a group photo from the number of face areas and perform smoothing processing suitable for the scene attribute.

  Further, for example, as shown in FIGS. 2A and 2B, the smaller the number of face regions, the larger the radius of the smoothing process, or the smoothing restriction is relaxed. It is possible to estimate a scene attribute of an image itself such as a non-person photograph, an identification photograph, a portrait, a snapshot, etc. from the number of face areas, and perform a smoothing process suitable for the scene attribute.

  As for the gradation contrast, for example, as shown in FIGS. 2C and 2D, the gradation contrast is shifted toward the high contrast side as the number of face regions increases. The larger the number of face areas, such as a group photo, from the number of face areas, the more the gradation contrast is shifted to the whole image is applied to the entire image, the scene attribute of the group photo or other image itself is estimated, and the more suitable for the scene attribute It is possible to perform a high gradation contrast.

  For example, as shown in FIGS. 2A and 2B, the smaller the number of face regions, the more the gradation contrast is shifted toward the soft side for the entire image. It is possible to estimate a scene attribute of an image itself such as a non-person photograph, an identification photograph, a portrait, a snap photograph, and the like from the number of face areas, and to perform gradation contrast softness suitable for the scene attribute.

  As for the defect removal processing, for example, as shown in FIGS. 2C and 2D, the defect removal processing is weakened or the defect removal processing is not applied as the number of face regions increases. It is possible to estimate a scene attribute of an image itself such as a group photo from the number of face areas and perform defect removal processing suitable for the scene attribute.

  For example, as shown in FIGS. 2A and 2B, the smaller the number of face regions, the larger the defect removal processing. It is possible to estimate a scene attribute of an image itself such as a non-person photograph, an identification photograph, a portrait, a snapshot, etc. from the number of face areas, and perform defect removal processing suitable for the scene attribute.

  An example of such image processing is shown in FIG.

  In this example, the image processing in the case where the number of detected face regions is 0, the sharpness enhancement degree is strong (high frequency enhancement), the contrast adjustment is high, and noise removal and defect removal are not applied. This image processing is performed under condition no. It was set to 6. In the case of a non-person photograph, such as a landscape photograph, an image having sharpness and high contrast was obtained.

  In the image processing when the number of detected face regions is 1, the sharpness enhancement degree is weak (low frequency enhancement), the contrast adjustment is soft, the noise removal is strong, and the defect removal is strong. This image processing is performed under condition no. It was set to 1. In the case of one person's ID photo and portrait, an image with low sharpness, high contrast at the face, and no scratches or dust was obtained.

  In the image processing when the number of detected face regions is 2 to 3, the sharpness enhancement degree is weak to medium (low frequency enhancement), contrast adjustment is not applied, noise removal is middle, and defect removal is standard. This image processing is performed under condition no. 2. In the case of a snapshot of 2 to 3 people, the sharpness was slightly weak, and an image with no noticeable scratches or dust was obtained.

  In the image processing when the number of detected face regions is 4 to 7, the sharpness enhancement degree is medium (low frequency enhancement), contrast adjustment is not applied, noise removal is middle, and defect removal is standard. This image processing is performed under condition no. It was set to 3. In the case of a group photo of 4 to 7 people, the sharpness was weak and an image with no noticeable scratches or dust was obtained.

  In the image processing when the number of detected face regions is 8 to 15, the sharpness enhancement degree is medium to strong (high frequency enhancement), contrast adjustment is not applied, noise removal is weak, and defect removal is weak. This image processing is performed under condition no. It was set to 4. In the case of a group photo of 8 to 15 people, the sharpness increased as the number of people increased, and an image without particularly noticeable scratches or dust was obtained.

  In the image processing when the number of detected face regions is 16 to 16, the sharpness enhancement degree is strong (high frequency enhancement), the contrast adjustment is hard, the noise removal is weak, and the defect removal is weak. This image processing is performed under condition no. It was set to 5. In the case of a group photo of 16 people or more, an image with enhanced sharpness as the number of people increased, and particularly without noticeable scratches or dust was obtained.

  In the image processing based on the image processing condition parameter, a face area having a smaller size than a predetermined face area is excluded from the number of face areas for image processing condition determination. For example, in the case of a group photo of an adult and a child, the size of the child's face area is smaller than a predetermined size, and this face area is excluded from the number of face areas for image processing condition determination. It is possible to estimate a scene attribute of an image such as a photograph and perform image processing suitable for the scene attribute.

  In addition, face areas with low high-frequency components are excluded from the number of face areas for image processing condition determination. This face area with a low high-frequency component is a blurred area, and by excluding such a blurred part from the number of face areas for image processing condition determination, non-person photos, ID photos, It is possible to estimate scene attributes of images such as portraits, snapshots, and group photos, and to perform image processing suitable for the scene attributes.

  Also, by changing the image processing conditions depending on the number of face areas and the size information of each face area, scenes of images such as non-person photos, ID photos, portraits, snapshots, group photos, etc. It is possible to estimate the attribute and perform image processing suitable for the scene attribute.

  An example of such image processing is shown in FIG.

  In this embodiment, when the number of detected face areas is 0, the image processing is as follows: face size reference × 1/4 or less, face size reference × 1/4 to 1/2, face size Reference Nos. 1/2 to 1, face size standards x 1 to 2, face size standards x 2 to 4, and face size standards x 4 or more are shown in condition No. 1 shown in FIG. 6 was processed. In the case of a non-person photograph, such as a landscape photograph, image processing was performed regardless of the size of the face, and an image having sharpness and high contrast was obtained.

  The image processing when the number of detected face areas is 1 is the condition No. shown in FIG. 3 in the case of the face size reference × 1/4 to 1/2 and the face size reference × 1/2 to 1, the condition No. 3 shown in FIG. 2, if the face size criterion is 1 to 2, the face size criterion is 2 to 4, and the face size criterion is 4 or more, the condition No. 2 shown in FIG. Processed by 1.

  Image processing when the number of detected face regions is 2 to 3 is as follows: face size reference × 1/4 or less, face size reference × 1/4 to 1/2, face size reference × 1 / 2 to 1, the condition No. shown in FIG. 3 in the case of face size standards x 1 to 2 and face size standards x 2 to 4, condition No. 3 shown in FIG. 3 is processed, and if the face size standard x 4 or more, the condition No. 3 shown in FIG. Processed by 1.

  The image processing when the number of detected face areas is 4 to 7 is the condition shown in FIG. 5 when the face size reference is ¼ or less, and when the face size reference is ¼ to 1/2. No. 4 in the case of the face size reference × 1 / 2˜1, the face size reference × 1-2, and the face size reference × 2-4, the condition No. 4 shown in FIG. 3 is processed, and if the face size standard x 4 or more, the condition No. 3 shown in FIG. Processed by 2.

  Image processing when the number of detected face areas is 8 to 15 is as follows: face size reference × 1/4 or less, face size reference × 1/4 to 1/2, face size reference × 1 / 2 to 1, the condition No. shown in FIG. 4 in the case of face size standards x 1 to 2 and face size standards x 2 to 4, condition no. 3 is processed, and if the face size standard x 4 or more, the condition No. 3 shown in FIG. Processed by 2.

  The image processing when the number of detected face areas is 16 to is the condition number shown in FIG. 5 when the face size reference is ¼ or less, and when the face size reference is ¼ to 1/2. . 5, and in the case of the face size reference × 1/2 to 1 and the face size reference × 1 to 2, the condition No. 5 shown in FIG. 4, when the face size standard x 2 to 4 and the face size standard x 4 or more, the condition No. 4 shown in FIG. Processed by 3.

  In this embodiment, a face area is detected by the face detection means, a face area of the maximum size is selected from the detected face areas, and an image processing condition parameter is generated based on the size information of the selected face area. Then, image processing is performed based on the image processing condition parameter. This image processing includes sharpness enhancement, contrast adjustment, noise removal, defect removal, edge enhancement, smoothing processing, and the like.

  In this way, the image processing condition parameter is generated based on the size information of the selected face area, and the image processing is performed based on the image processing condition parameter, so that the non-person photograph, the ID photograph, the portrait, the snap photograph, the set It is possible to estimate a scene attribute of an image such as a photograph and perform image processing suitable for the scene attribute.

  In this embodiment, the face area is detected by the face detection means, and the sharpness of each detected face area is determined. A face area having the maximum size is selected from those having a sharpness of a predetermined level or more, an image processing condition parameter is generated based on the size information of the selected face area, and image processing is performed based on the image processing condition parameter.

  This image processing includes sharpness enhancement, contrast adjustment, noise removal, defect removal, edge enhancement, smoothing processing, and the like.

  In this way, the face area having the maximum size is selected from those in which the sharpness of each detected face area is equal to or higher than a predetermined level, and image processing condition parameters are generated based on the size information of the selected face area. By performing image processing based on parameters, it is possible to estimate scene attributes of images such as non-person photos, ID photos, portraits, snapshots, group photos, etc., and perform image processing suitable for the scene attributes .

  This image processing method can be applied to image processing for recording an image on a recording material, a storage medium, an apparatus, or the like based on image data from a photographic original image or a digital camera, or displaying the image on a display device.

1 is a schematic configuration diagram of an image recording system that records an image on a recording material based on image data. It is a figure which shows an original image. It is a flowchart which shows an image process. It is a figure which shows sharpness emphasis. It is a figure which shows the Example of an image process. It is a figure which shows the Example of an image process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image recording system 2 Original image 3 Scanner 4 Image processing apparatus 5 Printer

Claims (14)

In an image processing method for recording or displaying an image based on image data,
The face area is detected by the face detection means,
Generating image processing condition parameters according to the number of detected face regions;
An image processing method, wherein image processing is performed based on the image processing condition parameter. The image processing method according to claim 1, wherein the degree of enhancement of a frequency component higher than a reference frequency for edge enhancement is relatively increased as the number of the face regions increases. The image processing method according to claim 1, wherein the degree of enhancement of frequency components lower than a reference frequency for edge enhancement is relatively increased as the number of face regions is smaller. The image processing method according to claim 1, wherein the larger the number of face regions, the smaller the radius of the smoothing process, or the non-application of the smoothing process. The image processing method according to claim 1, wherein as the number of face regions is smaller, a processing radius of the smoothing process is increased or a restriction on smoothing is loosened. The image processing method according to claim 1, wherein the larger the number of face regions, the more the image is applied with a gradation contrast shifted to the high contrast side. The image processing method according to claim 1, wherein the smaller the number of face regions, the more the image is applied with a gradation contrast shifted toward the soft side. The image processing method according to claim 1, wherein the defect removal processing is weakened or the defect removal processing is not applied as the number of the face regions increases. The image processing method according to claim 1, wherein the defect removal processing is increased as the number of the face regions is smaller. 11. The image according to claim 1, wherein a face area having a smaller size than the predetermined face area is excluded from the number of face areas for image processing condition determination. Processing method. 11. The image processing method according to claim 1, wherein a face area having a low high-frequency component is excluded from the number of face areas for image processing condition determination. 11. The image processing method according to claim 1, wherein the image processing condition is changed according to a combination of the number of the face areas and the size information of the face areas per one. 11. In an image processing method for recording or displaying an image based on image data,
The face area is detected by the face detection means,
Among the detected face areas, select the maximum size face area,
Generating an image processing condition parameter based on the size information of the selected face area;
An image processing method, wherein image processing is performed based on the image processing condition parameter. In an image processing method for recording or displaying an image based on image data,
The face area is detected by the face detection means,
Determine the sharpness of each detected face area;
Select the face area of the maximum size from those whose sharpness is a predetermined level or higher,
Generating an image processing condition parameter based on the size information of the selected face area;
An image processing method, wherein image processing is performed based on the image processing condition parameter.

Images