JP4645952B2 - Image identification method for ID photo and image processing apparatus - Google Patents

Description

  The present invention relates to an image processing technique for automatically performing correction according to the content (scene) of an original image, and in particular, a proof for automatically performing correction related to an ID photo image and obtaining a suitable output image. The present invention relates to a photographic image discrimination method and an image processing apparatus.

  In photography, there are various scenes, such as those with a landscape as a subject and mainly those with a person as a subject. In general, a photographic processing apparatus installed in a DPE store or the like has a function of performing correction (color density correction, contrast correction, etc.) according to each scene, and has a high quality suitable for the scene. Photo prints and the like are output and provided to customers.

  By the way, in recent years, with the spread of high-performance digital cameras, it is rare to use images taken with a camera that you own without going to a photo studio (including a speed photography box) as an image for an ID photo. Absent.

  For such an ID photo image, the corresponding correction (specifically, the contrast is weakened (compared to a landscape image), the gradation is corrected so that the gradation characteristics are soft, and the saturation is adjusted). It is generally known that an image that does not impair the texture of a person's skin, hair, or the like, that is, an image suitable for an ID photo can be obtained.

  As described above, a photo processing apparatus such as a DPE store can perform correction according to images of various scenes including such an image for ID photo, and a correction mode for a certain scene can be set in advance. For example, when an image is read from a recording medium or the like brought in by a customer, the correction related to the correction mode can be automatically performed.

  However, the automatic correction performs correction on the image according to a preset correction mode, and does not automatically determine the scene of the target image and perform correction according to the scene.

  Therefore, there are a lot of images brought in by customers, and when all scenes are mixed (in view of the current situation where the capacity of memory for digital cameras, etc. is increasing, this case is increasing) From the viewpoint of improving processing efficiency, normally, all images are automatically corrected in a preset correction mode (for example, landscape correction mode). In such a case, the correction works effectively on the landscape image, and a good-looking image is obtained. On the other hand, on the ID photo image, the texture of human skin, hair, etc. is adversely affected. There is a risk that.

  In this regard, it automatically detects whether a person's face is present in the original image (an image read from a photographic film or recording media brought in by the customer). There is also a photographic processing apparatus equipped with a function capable of automatically correcting such as. However, a method for reliably detecting whether or not it is for ID photo has not been established yet, so it is automatically determined whether or not the original image is for ID photo, and further, If it is a photographic image, there is no photographic processing apparatus (image processing apparatus) that can automatically perform corrections specialized for it.

  In view of the above problems, the present invention provides a new method that can reliably determine whether an original image is an image for an ID photo, and adopts such a determination method to Another object of the present invention is to provide a convenient image processing apparatus that can automatically perform correction on an ID photo image.

ID photo image determination method according to the present invention analyzes the original image performs detection of human face images, when the face image is detect, the detection number, orientation and the face image area of the face And a subject determination step for determining whether or not the subject related to the face image is suitable as a subject for an ID photo, and the subject determination step determines that the subject is suitable as a subject. A hue analysis step for analyzing a hue distribution of a background image, and when the hue distribution shows a predetermined tendency as a result of the analysis in the hue analysis step, the original image is used for an identification photograph. It is judged that it is an image.

  The present invention is characterized in that an image satisfying the two requirements of the subject requirement and the background requirement is judged to be appropriate as an image for ID photo. The subject requirement requires the validity of the certification photograph as the subject, and in the subject determination step, human face detection is performed from the original image, and various parameters (number of detections, face acquired) are acquired accordingly. And the size of the face image). Specifically, a face image of one person facing the front has a certain size (a size that allows the person to be identified, that is, a size that is appropriate as a proof photograph). In this case, it is determined that the subject requirement is satisfied (that is, the subject is suitable as an ID photo subject). For face detection (including acquisition of various parameters), a known face detection technique (provided by a software program or the like) can be employed.

  On the other hand, the background requirement requires the validity as a background image of the proof photo, and on the assumption that the subject requirement is satisfied, the hue distribution of the background image is analyzed in the hue analysis step. It is determined based on the analysis result. Specifically, when the hue distribution shows a predetermined tendency (in other words, when the hue distribution is biased to a specific hue, that is, when the background image is configured with a monotonous hue), the background image Is determined to be valid as the background image of the proof photo.

  As described above, according to the identification photo image discrimination method of the present invention, it is possible to discriminate whether or not the original image is an identification photo image by a simple method, so that the processing speed is high and the manufacturing is inexpensive. It is possible and excellent in practicality.

  In the hue analysis step, the background image of the upper region from the predetermined position in the vertical direction of the face in the original image is to be analyzed. For example, an image of clothes worn by the person is used as the background. Since it can be excluded from the image, the processing content can be further simplified without degrading the discrimination performance.

An image processing apparatus according to the present invention to which the ID photo image discrimination method is applied performs image processing on an original image and generates an output image. performs detection of facial images, when the face image is detect, the detection number, orientation and size of the face image area of the face as the determination parameter, the object of the subject related to the face image for identification photograph Subject determination means for determining whether or not the image is suitable, and hue analysis means for analyzing the hue distribution of the background image of the subject when the subject determination means is determined to be suitable for the subject. When the hue distribution shows a predetermined tendency as a result of the analysis by the means, the original image is corrected according to the ID photo image.

  Further, the hue analysis means is characterized in that the background image of the upper region from the predetermined position in the vertical direction of the face in the original image is an analysis target.

  According to the image processing apparatus having the above configuration, it is automatically determined whether or not the original image read from the photographic film or the recording medium is an ID photo image. The correction process can be performed automatically, so even if multiple image data (image files) contain image data captured for use in ID photos, the ID photo images are automatically identified. Therefore, appropriate correction can be performed.

  As described above, according to the ID photo image discrimination method of the present invention, since it is possible to realize whether or not the original image is an ID photo image by a simple method, for example, many images captured in a plurality of scenes are used. It is possible to reliably and quickly select an ID photo image from among the images.

  Further, according to the image processing apparatus of the present invention, it is possible to automatically determine whether or not the read original image is an ID photo image, and if it is an ID photo image, a correction process specialized for it is performed. It can be done automatically. Therefore, for example, when the image processing apparatus is applied to a photo processing apparatus, even if ID image is mixed in many image data (image files) captured in a plurality of scenes brought by a customer, Appropriate correction can be automatically applied only to the ID photo image, and the service to customers can be expanded.

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

  FIG. 1 is an external view of a photographic processing apparatus provided with an image processing apparatus according to the present embodiment, and FIG. 2 is a block diagram showing an internal configuration of the photographic processing apparatus. This photographic processing apparatus reads image data from a photographic film F or a recording medium M, performs digital image processing such as color density correction and contrast correction on the image data, and generates image data for output. 1 and a printing apparatus 2 that prints an image on the photosensitive material PA based on the image data output from the image processing apparatus 1.

  The image processing apparatus 1 includes a film scanner 3 that reads an image from a photographic film F, a media drive 4a that can read and write image data from and to a recording medium M such as a memory card and a CD-R, and a large number of images read from the recording medium M. Large-capacity hard disk 4b capable of recording the image data (image file) and the like, a main control device 5 for controlling the entire photographic processing device including the control of the image processing device 1, and operations related to the control of the main control device 5 A monitor 6 for displaying a screen and the like and an input device 7 (a keyboard 7a and a mouse 7b) for receiving an operation command from an operator are provided. Note that the monitor 6 may be configured by a touch panel method, and an operation command may be received using an operator's finger, a dedicated input pen, or the like.

  The main control device 5 includes a CPU 8 that is a main body of the control and calculation, a ROM 9a that stores an image processing program, a RAM 9b that is used as a work area for processing, and various external devices (film scanner 3, media drive 4a). The hard disk 4b, the monitor 6 and the input device 7) are provided with an I / O processing unit 10 for performing I / O processing.

  The printing apparatus 2 includes an exposure processing unit 11 that performs exposure processing on the photosensitive material PA based on image data that has been subjected to image processing by the image processing device 1, and a development processing unit 12 that performs development processing on the exposed photosensitive material PA. The drying processing unit 13 performs drying processing on the developed photosensitive material PA, and the sorting processing unit 14 performs sorting processing on the photosensitive material PA that has been dried and turned into a photograph.

  The exposure processing unit 11 pulls out the photosensitive material PA from the pair of magazines 15 a and 15 b and conveys it to the downstream side by the conveying roller pair 16. Then, the photosensitive material PA is exposed by a PLZT type exposure device 17 including a light source 17a composed of a halogen lamp or the like and an optical shutter 17c connected to the light source 17a via an optical fiber bundle 17b. The exposure device 17 is not limited to the PLZT method, and may be a laser method, a CRT method, or the like.

  The development processing unit 12 performs processing such as development, bleaching, fixing, and stabilization on the photosensitive material PA.

  The drying processing unit 13 performs a process of blowing hot air from the hot air device 18 to the photosensitive material PA sent from the development processing unit 12 to dry the phenomenon liquid adhering to the photosensitive material PA.

  The sorting processing unit 14 sorts the photosensitive material PA that has been subjected to the exposure processing, development processing, and drying processing according to size, and the small-sized photosensitive material PA sorted by the size sorting unit 19 for each order. And an order sorting unit 20 for sorting.

  The size sorting unit 19 discharges the photosensitive material PA having a large size to the tray 21 and conveys the photosensitive material PA having a small size to the conveyor 22. The order sorting unit 20 distributes the photosensitive material PA conveyed by the conveyor 22 to each tray 23 for each order.

  In addition to the above general functions, the photographic processing apparatus according to the present embodiment determines whether or not the original image data read from the photographic film F or the recording medium M brought in by the customer is image data for an ID photo. It has an unprecedented superior feature that it can automatically determine and automatically perform a correction process specialized for it. Hereinafter, this feature will be described in detail.

  FIG. 3 is a flowchart showing a flow of identification photo image data discrimination processing performed by the image processing apparatus 1. Here, the processing function unit related to the determination process of the image processing apparatus 1 (specifically, the main control apparatus 5) constitutes a subject determination unit and a hue analysis unit. First, when original image data is read from the photographic film F or the recording medium M, a process (face detection process) for detecting a human face from the original image is performed (step S301). As face detection processing here, a method in which a region where a specific color (for example, skin color) continues is used as a face region, a specific shape pattern using Hough transform or the like, a structure extraction method by template matching Alternatively, a known technique such as a method combining these may be employed.

  If no human face is detected (No in step S302), the process ends. In this case, since it is not regarded as image data for ID photos, automatic correction is performed in accordance with a preset correction mode (for example, a landscape correction mode (high contrast and high contrast correction), etc.). . On the other hand, if a face is detected (Yes in step S302), it is checked whether or not the number of detection is one (that is, one person as a subject), and if the number of detection is plural (step S303). No), this processing ends, and correction processing is executed according to a preset correction mode as described above. On the other hand, when the number of detection is one (Yes in step S303), it is checked whether or not the detected face is facing the front. Here, for the determination of the face direction, any known technique can be adopted. However, in order to avoid complication of processing, for example, when both eyes are detected, the specification may be that the front faces. .

  As a result of the determination of the face direction, when it is determined that it is not facing the front (No in step S304), this processing ends, and correction processing is executed according to a preset correction mode. On the other hand, if it is determined that it is facing the front (Yes in step S304), the validity of the size of the face image is further checked. Specifically, the ratio of the face image area to the entire original image is calculated, and it is determined whether or not the ratio has validity for the ID photo.

  As a result, when it is determined that the size is not appropriate (No in step S305), the process ends, and the correction process is executed according to a preset correction mode. On the other hand, when it is determined that the size is appropriate (Yes in step S305), it is determined that the subject of the original image is appropriate as the subject of the certification photograph. As a specific example, the image data of only the landscape as shown in (b) of FIG. 4 (no person is captured) is excluded by the determination in step S302, and a plurality of images as shown in (c) of FIG. Image data showing a person is excluded by the determination in step S303, and image data in which the face of the subject person as shown in FIG. 4D is not facing the front is excluded by the determination in step S304. Furthermore, image data with a small face image area as shown in FIG. 4E is excluded by the determination in step S305.

  When the validity of the subject is confirmed by the above processing, the background image data analysis processing is subsequently performed (step S306). In this analysis processing, the hue of each pixel constituting the background image data (that is, the image data of the original image data excluding the image data portion related to the person) is obtained, and distribution analysis is performed on this. In this case, only the background image data of the upper region from the predetermined position in the vertical direction of the face in the original image data is analyzed in order to simplify the process by excluding the image data related to the clothes of the person from the background image data. The target specification may be used. FIG. 5 shows an example in which the predetermined position in the vertical direction of the face is in the vicinity of the lips of the subject (person). In this case, the region above the one-dot chain line (the region indicated by the oblique lines) in the figure is the background image data to be analyzed. The hue of each pixel is converted from RGB original image data to YCbCr format data (data composed of lightness (Y) and saturation (C) components) using, for example, a predetermined conversion table. From the converted Cb and Cr, it can be obtained using a predetermined calculation formula or the like.

  As a result of the above analysis, when it is not recognized that the background image data is biased to a specific hue from the hue distribution (No in step S307), the original image data is not regarded as image data for an ID photo, This process ends, and the correction process is executed in accordance with a preset correction mode. With this determination, for example, as shown in FIG. 4F, image data whose background data is not suitable for the ID photo is surely excluded although the requirements as a subject are satisfied.

  On the other hand, when it is recognized that the background image data is biased to a specific color (Yes in step S307), the original image data is regarded as image data for an ID photo (see (A) in FIG. 4). Regardless of the correction mode set in advance, the predetermined correction relating to the ID photo image is performed (step S308). Specifically, correction is performed such that the contrast is weak and the gradation characteristics are soft.

  As described above, according to the image processing apparatus 1 of the present embodiment of the present invention, it is automatically determined whether or not the original image data read from the photographic film or the recording medium is the image data for the ID photo. If it is determined that the image data is for ID photo, correction processing specialized for it can be automatically performed. Therefore, image data suitable for ID photo (that is, texture of human skin and hair) Image data that does not impair the image quality). In addition, since the determination function employed in the image processing apparatus 1 can be realized by simple processing, the processing speed is high and the manufacturing cost is low.

  In addition, even if the original image data that the customer did not intend for the ID photo is determined to be image data for the ID photo by the image processing apparatus 1, and correction processing is performed accordingly. Therefore, it can be said that there is little possibility of causing a disadvantage to the customer and it is excellent in practicality.

  In addition, this invention is not limited to the said embodiment, Of course, a various change is possible in the range which does not deviate from the summary of this invention.

1 is an external view of a photographic processing apparatus according to an embodiment of the present invention. It is a schematic block diagram of the photographic processing apparatus which concerns on the same embodiment. 4 is a flowchart showing a flow of processing for determining whether or not the image data is ID photo data in the embodiment. In the same embodiment, it is a figure which shows the example of an image for demonstrating whether it is suitable as an object for ID photographs. In the embodiment, it is a figure for demonstrating an example for limiting the area | region of the background image data made into analysis object.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image processing apparatus, 2 ... Printing apparatus, 3 ... Film scanner, 4a ... Media drive, 4b ... Hard disk, 5 ... Main controller, 6 ... Monitor, 7 ... Input device, 8 ... CPU, 9a ... ROM, 9b ... RAM, 10 ... I / O processing unit, 11 ... exposure processing unit, 12 ... development processing unit, 13 ... drying processing unit, 14 ... sorting processing unit, 15a, 15b ... magazine, 16 ... conveying roller pair, 17 ... exposure 17a ... light source, 17b ... optical shutter, 18 ... hot air device, 19 ... size sorting unit, 20 ... order sorting unit, 21 ... tray, 22 ... conveyor, 23 ... dish

Claims (4)

Analyzes the original image performs detection of human face images, when the face image is detect, the detection number, orientation and size of the face image area of the face as the determination parameter, to the face image A subject determination step for determining whether or not the subject is suitable as a subject for the ID photo;
A hue analysis step for analyzing a hue distribution of a background image of the subject when the subject determination step determines that the subject is suitable as a subject,
As a result of the analysis in the hue analysis step, when the hue distribution shows a predetermined tendency, the original image is determined to be an image for an ID photo.   2. The ID photo image discrimination method according to claim 1, wherein in the hue analysis step, a background image of an upper region from a predetermined position in the vertical direction of the face in the original image is an analysis target. In an image processing apparatus that performs image processing on an original image and generates an output image,
Performs detection of human face images by analyzing the read original image, when the face image is detect, the detection number, orientation and size of the face image area of the face as the determination parameter, the face Subject determination means for determining whether or not the subject related to the image is suitable as a subject for the ID photo;
Hue analysis means for analyzing the hue distribution of the background image of the subject when the subject determination means is determined to be suitable as a subject,
As a result of the analysis by the hue analysis means, if the hue distribution shows a predetermined tendency, the original image is corrected according to the ID photo image.   The image processing apparatus according to claim 3, wherein the hue analysis unit analyzes a background image of an upper region from a predetermined position in the vertical direction of the face in the original image.

Images