JP5332668B2 - Imaging apparatus and subject detection program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an imaging device and a subject detection program that can make processings perform rapidly for subject detection, using an easy method. <P>SOLUTION: The imaging device (11 to 26) detects a subject from an image, by using predetermined parameters of subject detection and includes: an attitude information acquisition means (21, 20) of acquiring information showing the attitude of the device; and a subject detection means (20) of changing the parameters of subject detection according to the attitude of the device, in which the information acquired by the attitude information acquisition means shows, when the subject is detected from the image. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to an imaging device that automatically detects a subject from an image and a subject detection program.

  Many methods for automatically detecting a subject from an image have been proposed.

  By the way, due to the recent increase in resolution of images, it takes time to detect the subject automatically in detecting the subject from the image. For this reason, when shooting with a digital camera or the like having a subject detection function, problems such as a large release time lag and a slow continuous shooting speed have occurred.

  As a technique for solving such a problem, for example, in Patent Document 1, face detection (subject detection) is performed by changing a detection method for detecting a face area of a subject from an image according to the posture of a digital camera or the like. Discloses a method of speeding up the process.

JP 2005-130468 A

  By the way, in general, a portion where an object (particularly a face) is captured in an image is generally a specific region corresponding to the posture of the camera in the image. For example, when shooting is performed with the camera in a normal landscape orientation, the captured image includes a subject (particularly a face) as shown in FIG. 9A unless the subject is standing upside down. In many cases, the image is shown in an area with the center of gravity placed above the image. In many cases, the direction of the subject (particularly, the face) in the image is a specific direction corresponding to the posture of the camera. For example, when shooting is performed with the camera in a normal sideways posture, the captured image has a top of the subject, as shown in FIG. 9B. 1) When lying on the side, it is often reflected in the shape facing left (4) or right (2), and the top of the head is below (3) as if the subject is standing upside down. It is rare that it is reflected in the form facing the.

  That is, the part (area) in which the subject (particularly the face) in the image is shown, and the orientation (direction) of the subject in the image can be specified by the posture of the camera.

  However, the conventional technique changes the detection method of the face area (subject) according to the posture of the camera, but when detecting the face area of the subject from the image, the entire area in the image is detected. Detection processing was performed. Further, the conventional technology does not detect the face area of the subject from the image in consideration of the direction of the subject in the image.

  For this reason, there is room for improvement in speeding up the face detection (subject detection) process.

  Also, if the detection method is changed according to the camera posture as in the conventional technology, if there is an additional type of posture supported by the camera, it is also necessary to add a new detection method corresponding to it. It was not possible to cope with the addition easily.

  For this reason, it is necessary to realize a high-speed face detection (subject detection) process by a simpler method than changing the detection method.

  The present invention is to solve the above-mentioned problems of the prior art. An object of the present invention is to provide an imaging apparatus and a subject detection program capable of speeding up subject detection processing by a simple method.

An image pickup apparatus according to a first aspect of the present invention is an image pickup apparatus that detects a subject from an image using predetermined subject detection parameters, and includes posture information acquisition means for acquiring information indicating the posture of the device; when performing the detection of the subject from an image, according to the posture of the apparatus shown has been acquired information by the attitude information acquisition means, and a subject detecting means for changing the parameters of the object detection, object detection means, orientation obtaining Information indicating the posture is acquired by the means, and a reduced image is created by reducing the image size to a predetermined size based on the image in parallel with changing the subject detection parameter according to the acquired posture. The subject is detected using the created reduced image .

  In a second aspect based on the first aspect, at least one of the subject detection parameters is a detection target region that is a region in the image where the subject is detected and a detection non-target region where the subject is not detected. It is a parameter that specifies the category.

  According to a third aspect, in the first or second aspect, at least one of the subject detection parameters is a parameter for designating a pattern of a subject direction for which detection is effective in an image.

  In a fourth aspect based on any one of the first to third aspects, the subject detection means always detects a subject using an image in a lateral position regardless of the posture of the apparatus.

A subject detection program according to a fifth aspect of the present invention is a subject detection program for detecting a subject from an image using a predetermined subject detection parameter, and a posture information acquisition step for acquiring information indicating the posture of the apparatus If, when performing detection of the subject from an image, depending on the orientation of the device indicated by the information acquired by the orientation information acquiring step, to execute the subject detection step of changing parameters of the object detection in a computer, the subject detection step In the posture information acquisition step, information indicating the posture is acquired, and the image size is reduced to a predetermined size based on the image in parallel with changing the subject detection parameter according to the acquired posture. A reduced image is created, and a subject is detected using the created reduced image .

In a sixth aspect based on the fifth aspect , at least one of the subject detection parameters is a detection target region that is a region where the subject is detected in the image and a detection non-target region where the subject is not detected. It is a parameter that specifies the category.

In a seventh aspect based on the fifth or sixth aspect , at least one of the subject detection parameters is a parameter for designating a pattern in the direction of the subject for which detection is effective in the image.

Eighth invention, in any one invention of the fifth to seventh invention, in the subject detecting step, regardless of the orientation of the device, the detection of an object always use the image of the lateral position.

  In the present invention, when subject detection is performed from an image, subject detection parameters are changed according to the orientation of the device indicated by the acquired information.

  Therefore, by using the present invention, it is possible to speed up the subject detection process by a simple method that only changes the subject detection parameters.

It is a block diagram of the digital camera of this embodiment. It is a figure explaining the attitude | position and attitude | position information of a digital camera. It is a flowchart which shows the operation | movement of the face detection of a digital camera. It is a figure which shows the image (horizontal position) of the image image | photographed for every attitude | position. It is a figure explaining a "non-detection mask". It is a figure explaining "direction pattern". It is a flowchart (modification) which shows operation of face detection of a digital camera. It is a figure which shows the image (vertical position) of the image image | photographed for every attitude | position. It is a figure which shows the general example of the area | region and the direction of a to-be-photographed object in the image.

  Embodiments of the present invention will be described below. The present embodiment is an embodiment of a digital camera.

  FIG. 1 is a block diagram of the digital camera of this embodiment.

  The digital camera includes an imaging lens 11 and a lens driving unit 12, an imaging element 13, an analog signal processing unit 14, a timing generator (TG) 15, a buffer memory 16, an image processing unit 17, and a display control unit 18. A display unit 19, a control unit 20, a sensor unit 21, a compression / decoding unit 22, a recording interface (recording I / F) 23, a recording medium 24, an operation unit 25, and a bus 26. ing. Here, the buffer memory 16, the image processing unit 17, the display control unit 18, the control unit 20, the sensor unit 21, the compression / decoding unit 22, and the recording I / F 23 are connected via a bus 26. Further, the lens driving unit 12, the analog signal processing unit 14, the TG 15, and the operation unit 25 are connected to the control unit 20, respectively.

  The imaging lens 11 includes a plurality of lens groups including a focus lens and a zoom lens. For simplicity, the imaging lens 11 is illustrated as a single lens in FIG.

  The lens driving unit 12 generates a lens driving signal in accordance with an instruction from the control unit 20, performs focus adjustment and zoom adjustment by moving the imaging lens 11 in the optical axis direction, and subjects the light flux that has passed through the imaging lens 11 to the subject. An image is formed on the light receiving surface of the image sensor 13.

  The imaging element 13 is a CCD type or CMOS type imaging element, and is arranged on the image space side of the imaging lens 11. The image sensor 13 photoelectrically converts a subject image formed on the light receiving surface to generate an analog image signal. The output of the image sensor 13 is connected to an analog signal processing unit 14.

  The analog signal processing unit 14 performs analog signal processing such as CDS (correlated double sampling), gain adjustment, and A / D conversion on the analog image signal output from the image sensor 13 in accordance with an instruction from the control unit 20. And the processed image signal is output. The output of the analog signal processing unit 14 is connected to the buffer memory 16.

  Further, the analog signal processing unit 14 sets an adjustment amount for gain adjustment based on an instruction from the control unit 20, and thereby adjusts photographing sensitivity corresponding to ISO sensitivity.

  The TG 15 supplies timing pulses to the image sensor 13 and the analog signal processing unit 14 based on instructions from the control unit 20. The drive timing of the image sensor 13 and the analog signal processing unit 14 is controlled by the timing pulse.

  The buffer memory 16 temporarily stores the image signal output from the analog signal processing unit 14 as image data. The buffer memory 16 temporarily stores image data read from the recording medium 24 by the control unit 20 and image data created in the course of processing by the control unit 20.

  The image processing unit 17 performs image processing such as white balance adjustment, interpolation, contour enhancement, gamma correction, resolution conversion (pixel number conversion) on the image data in the buffer memory 16 in accordance with an instruction from the control unit 20. Note that resolution conversion (pixel number conversion) is a process necessary for displaying image data on the display unit 19. The image processing unit 17 is configured as an ASIC or the like.

  In response to an instruction from the control unit 20, the display control unit 18 performs predetermined signal processing (for example, signal processing for converting into an NTSC composite video signal) on the image data after the image processing, and displays the image data on the display unit 19. Output. With this output, the image data is displayed on the display unit 19. The display unit 19 is an LCD monitor provided on the back surface of the digital camera housing, an electronic viewfinder having an eyepiece unit, or the like.

  The sensor unit 21 includes a sensor for detecting the attitude of the digital camera, for example, a tilt sensor. The control unit 20 drives the sensor unit 21 to acquire information (posture information) indicating the posture of the digital camera.

  Here, the posture information will be described with reference to FIG. In the following description, the surface on which the LCD monitor screen is provided is described as the front of the digital camera. In the present embodiment, when the bottom surface of the digital camera is in the direction of gravity as shown in FIG. 2A, for example, posture information “lateral” is acquired. Further, as illustrated in FIG. 2B, when the left side surface of the digital camera is in the gravity direction, for example, posture information “vertical −90 degrees” is acquired. In addition, when the right side surface of the digital camera is in the direction of gravity as illustrated in FIG. 2C, for example, posture information “vertical +90 degrees” is acquired. As described above, in the digital camera according to the present embodiment, three types of posture information of “horizontal”, “vertical −90 degrees”, and “vertical +90 degrees” are acquired. For convenience of explanation, “horizontal”, “vertical −90 degrees”, and “vertical +90 degrees” of the posture information are used as they are as names for representing the posture of the digital camera.

  The compression / decoding unit 22 performs compression processing or decoding processing on the image data in the buffer memory 16 in accordance with an instruction from the control unit 20. The compression process and the decoding process are performed in a JPEG (Joint Photographic Experts Group) format or the like. The compression / decoding unit 22 is configured to be able to perform lossless compression (so-called lossless encoding).

  The recording I / F 23 is formed with a connector for connecting the recording medium 24. The recording I / F 23 accesses the recording medium 24 connected to the connector, and writes and reads various data. At the time of shooting, the control unit 20 records the image data after the compression processing of the buffer memory 16 on the recording medium 24 via the recording I / F 23. However, when the digital camera is set to the uncompressed recording mode, the control unit 20 records the image data after the image processing in the buffer memory 16 on the recording medium 24 without being compressed, without being compressed. . Further, the controller 20 reads the recorded image data from the recording medium 24 via the recording I / F 23 and records it in the buffer memory 16 during image reproduction. The recording medium 24 is a memory card with a built-in semiconductor memory, a small hard disk, or the like.

  The operation unit 25 includes various operation members such as a release button (shutter button), a mode setting button, a menu button, an image reproduction button, and the like, and an operation signal corresponding to the content of the member operation by a user such as a photographer. Send to.

  The control unit 20 comprehensively controls each unit of the digital camera according to the operation content of the operation member by the user.

  For example, when the digital camera is set to the shooting mode by operating the mode setting button or the like, the control unit 20 drives the lens driving unit 12, the analog signal processing unit 14, and the TG 15 to start shooting a through image. At this time, the image sensor 13 is driven in the draft mode (decimation readout mode), and the image data of the through image is sequentially recorded in the buffer memory 16 via the analog signal processing unit 14. The control unit 20 performs focus adjustment control (AF) of the imaging lens 11 in cooperation with the lens driving unit 12 based on the image data of the through image. In particular, as a face recognition function of a digital camera, for example, when face recognition AF is enabled, the control unit 20 detects a face portion (face area) of a subject from image data of a through image, and Focus adjustment control (AF) is performed so that the face area is in focus. Further, when the control unit 20 drives a photometry unit (not shown) to calculate the evaluation value of the shooting scene based on the image data of the through image in the buffer memory 16, the analog signal processing unit 14 and the like are based on the evaluation value. Adjust the settings of. The control unit 20 drives the image processing unit 17 to perform image processing on the image data of the through image in the buffer memory 16, and then drives the display control unit 18 to display the image data of the through image after image processing. The information is sequentially displayed on the display unit 19 (live view display).

  Next, when the release button is pressed halfway during operation in the shooting mode, the control unit 20 performs focus adjustment control (AF) of the imaging lens 11 in cooperation with the lens driving unit 12 prior to the main shooting. At this time, if the face area of the subject is detected, focus adjustment control (AF) is performed so that the face area is in focus.

  When the release button is fully pressed during operation in the shooting mode, the control unit 20 sets shooting conditions (aperture value, shutter speed, presence / absence of flash emission, etc.) based on an evaluation value calculated by a photometry unit (not shown). decide. Then, the control unit 20 drives the lens driving unit 12, the analog signal processing unit 14, and the TG 15 under the determined shooting conditions to perform the main shooting. At this time, the image sensor 13 is driven in a frame mode (all pixel readout mode), and image data of a captured image (main captured image) is recorded in the buffer memory 16 via the analog signal processing unit 14. Thereafter, the control unit 20 drives the image processing unit 17 to perform image processing on the image data of the main captured image recorded in the buffer memory 16. In addition, the control unit 20 drives the display control unit 18 to display the actual captured image after image processing on the display unit 19. Then, the control unit 20 drives the compression / decoding unit 22 to perform compression processing on the image data of the actual captured image after the image processing, and record the image data of the actual captured image after the compression processing. To the recording medium 24. When the digital camera is set to the uncompressed recording mode, the control unit 20 records the image data of the actual captured image after image processing without being compressed without driving the compression / decoding unit 22. Recording is performed on the recording medium 24 via / F23.

  When the image reproduction button is pressed, the control unit 20 reads out image data such as a main photographic image recorded on the recording medium 24 to the buffer memory 16 as image data of the reproduced image via the recording I / F 23. . Thereafter, the control unit 20 drives the compression / decoding unit 22 as necessary to perform decoding processing on the image data in the buffer memory 16. As a face recognition function of the digital camera, for example, when face recognition / playback is enabled, the control unit 20 detects the face area of the subject from the image data in the buffer memory 16. Further, the control unit 20 drives the image processing unit 17 as necessary, and performs necessary image processing on the image data in the buffer memory 16. Then, the control unit 20 drives the display control unit 18 to display the image data in the buffer memory 16 on the display unit 19. At this time, if the face area of the subject is detected, the display unit 19 displays an image in which a frame (detection frame) indicating the detection is superimposed on the face area, or centered on the face area. An enlarged image is displayed.

  Hereinafter, the face detection operation of the digital camera will be described with reference to the flowchart of FIG. FIG. 3 is a flowchart of face detection processing executed during shooting and image playback operations when the digital camera face recognition function such as face recognition AF and face recognition playback is enabled.

  Step 101 (S101): The control unit 20 converts any one of the through image, the actual captured image, and the reproduced image recorded in the buffer memory 16 by the photographing and image reproduction operations into the target image (for the face detection process). To be used as a detection target image).

  In the digital camera of the present embodiment, when the posture is “horizontal”, as shown in FIG. 4A, an image of “horizontal position” displayed on the LCD monitor screen is captured and buffered. Recorded in the memory 16. When the posture is “vertical-90 degrees”, as shown in FIG. 4B, an image obtained by rotating the image displayed on the LCD monitor screen 90 degrees clockwise (clockwise) is photographed. Recorded in the buffer memory 16. When the posture is “vertical +90 degrees”, an image obtained by rotating the image displayed on the LCD monitor screen 90 degrees counterclockwise (counterclockwise) as shown in FIG. Recorded in the buffer memory 16. That is, in the digital camera according to the present embodiment, these “lateral position” images recorded in the buffer memory 16 are used as detection target images regardless of the posture.

  Step 102: The control unit 20 drives the sensor unit 21 to acquire information (posture information) indicating the posture of the digital camera. However, when the processing of this flowchart is executed by an image reproduction operation such as face recognition reproduction, the control unit 20 acquires posture information from a reproduction image file. If the reproduction image file is based on the Exif format, the posture information is recorded as the tag information.

  Step 103: The control unit 20 determines face detection parameters according to the posture of the digital camera indicated by the obtained posture information.

  Here, face detection parameters will be described. First, the “non-detection mask” that is the first face detection parameter will be described with reference to FIG.

  The “non-detection mask” is a parameter for designating a classification of a region (detection target region) where face detection processing is performed in a detection target image and a region where detection processing is not performed (detection non-target region). In “Non-detection mask 1” to “Non-detection mask 3” shown in FIG. 5, a detection non-target region is specified by a hatched region, and a detection target region is specified by a white region. To do.

  In the digital camera according to the present embodiment, when the posture is “horizontal”, as shown in FIG. 5A, “non-display” for designating a region having the center of gravity above the detection target image as the detection target region. A detection mask 1 ”is used. When the orientation is “vertical-90 degrees”, as shown in FIG. 5B, a “non-detection mask 2” for designating a region with the center of gravity on the right side of the detection target image as the detection target region. Is used. Further, when the posture is “vertical +90 degrees”, as shown in FIG. 5C, “non-detection mask 3” for designating a region having the center of gravity on the left side of the detection target image as the detection target region. Is used.

  In other words, the “non-detection mask” means that the face of the subject is usually present in the portion taken from the general composition in the images taken in the postures of “horizontal”, “vertical −90 degrees”, and “vertical +90 degrees”. This is for designating a region considered to be nonexistent as a non-detection target region. For example, if the face of the subject is unlikely to be present in the lower part of the image taken in the “landscape” position unless the camera is upside down, the digital camera is in the “landscape” position. For this, the “non-detection mask 1” shown in FIG. 5A is used to cover the detection target image. As a result, an area in which the face in the detection target image is considered not to exist is designated as a non-detection target area, and face detection processing is not performed for that area. Therefore, face detection can be performed at a higher speed than when face detection processing is performed on the entire region of the detection target image.

  The information of “non-detection mask 1” to “non-detection mask 3” is stored in advance in a storage medium such as a ROM (not shown) of the digital camera.

  From the above, the control unit 20 sets the “non-detection mask 1” when the acquired posture information is “horizontal”, and the “non-detection mask 2” when the posture information is “vertical-90 degrees”. When the information is “vertical +90 degrees”, “non-detection mask 3” is determined as the first face detection parameter.

  Next, the “direction pattern” that is the second face detection parameter will be described with reference to FIG.

  The “direction pattern” is a parameter for designating the pattern of the face direction (direction) of the subject for which detection is effective in the detection target image.

  In the digital camera of this embodiment, the faces in four directions shown in (1) to (4) of FIG. 6A can be detected from the detection target image. In the digital camera of the present embodiment, as shown in FIG. 6B, when the posture is “horizontal”, the faces in the directions shown in (4), (1), and (2) among the four “Direction pattern 1” is used to specify that the detection of the above is valid. Further, when the posture is “vertical-90 degrees”, a “direction pattern” for designating that the detection of the face in the directions shown in (1), (2), and (3) is valid among the four. 2 "is used. Further, when the posture is “vertical +90 degrees”, “direction pattern 3” for designating that the detection of the face in the directions shown in (1), (4), and (3) is valid among the four. Is used.

  In other words, it is considered that the “direction pattern” is usually captured in the direction from the general composition in the detection target image taken in the postures of “horizontal”, “vertical −90 degrees”, and “vertical +90 degrees”. This is for designating the face direction (at least one or more) as a detection target. For example, if the detection target image is taken in a posture of “vertical +90 degrees”, the face in the direction shown in (2) of FIG. 6A is not captured unless the subject is standing upside down. it is conceivable that. Therefore, in the digital camera of this embodiment, when the posture is “vertical +90 degrees”, the face in the direction (1), (4), (3) except (2) in FIG. “Direction pattern 3” in FIG. 6B, which specifies that detection is valid, is used. As a result, the face detection process is not performed for a face in a direction considered not to be captured. Therefore, the face detection can be performed at a higher speed than when face detection processing is performed on faces in all directions shown in (1) to (4) of FIG.

  In the “direction pattern 1” to “direction pattern 3” shown in FIG. 6B, the face detection in three directions is specified to be valid, but the “direction pattern” is specified. The number of face directions is not limited to three, and any number of face directions may be specified. That is, only one may be specified, or all (four in FIG. 6A) may be specified. For example, when the posture of the digital camera is “horizontal”, it is considered that the direction of (1) in FIG. 6 (a) usually needs to be designated as a “direction pattern” as the face direction. The directions of (4) and (2) in a) are for cases where the subject is lying on the side, so if these are excluded from face detection targets, these two directions are It is not necessary to specify with "Pattern". Therefore, when these two directions are deleted from “direction pattern 1” in FIG. 6B, the face direction in which face detection is performed when the posture of the digital camera is “lateral” is (1) in FIG. ) In only one direction, face detection can be performed at higher speed.

  The information of “direction pattern 1” to “direction pattern 3” is stored in advance in a storage medium such as a ROM (not shown) of the digital camera.

  From the above, the control unit 20 displays “direction pattern 1” when the acquired posture information is “horizontal”, “direction pattern 2” when the posture information is “vertical-90 degrees”, and the posture information is In the case of “vertical +90 degrees”, “direction pattern 3” is determined as the second face detection parameter.

  Step 104: The control unit 20 detects the face area of the subject from the detection target image using the face detection parameter determined as described above.

  That is, after the “non-detection mask” determined as the first face detection parameter is placed on the detection target image, the face area of the subject is detected from the detection target area of the “non-detection mask”. At this time, the detection of the face area of the subject is performed only for the face in the direction in which the detection is valid based on the “direction pattern” determined as the second face detection parameter.

  Note that the detection of the face area of the subject may be performed using only the first face detection parameter, and of course, it may be performed using only the second face detection parameter.

  The detection of the face area of the subject may be performed using a conventional “pattern matching method” disclosed in, for example, Japanese Patent Application Laid-Open No. 2006-1555096.

(Supplementary items of the embodiment)
In the above description, in step 101, the image data recorded in the buffer memory 16 by the shooting and image reproduction operations is set to be used as it is as the target image of the face detection process, that is, the detection target image. However, in order to perform face detection at high speed, for example, as shown in the flowchart of FIG. 7, a part of the face detection process of the flowchart of FIG. 3 may be modified. Specifically, as shown in the flowchart of FIG. 7, in step 101, based on the image data recorded in the buffer memory 16, a reduced image is created by reducing the image size to a predetermined size. Then, at the beginning of the processing in step 104, the reduced image is set to be used as a detection target image. In this way, if face detection is performed using a reduced image, the amount of data in pattern search in face detection processing is reduced compared to the case of using an image having an original large image size. The time required for face detection can be shortened. However, in such a case, the information on the “non-detection mask” used as the first face detection parameter is prepared in advance according to the image size of the reduced image and stored in the ROM (not shown) of the digital camera. Store in advance on the medium.

  In the flowchart of FIG. 7, processing for creating a reduced image in step 101, posture detection in step 102, and face detection parameter determination in step 103 so as not to increase the time required for face detection by creating a reduced image. The process is executed in parallel. In that case, for example, the control unit 20 drives the image processing unit 17 to cause the image processing unit 17 to create a reduced image. At the same time, the control unit 20 drives the sensor unit 21 to acquire posture information, and determines a face detection parameter according to the posture of the digital camera indicated by the posture information.

  Further, a predetermined area where the subject's face is considered to exist may be cut out from the image data recorded in the buffer memory 16, and the cut-out image (partial image) may be used as the detection target image. By doing so, compared to the case where the image data in the buffer memory 16 is used as a detection target image, the pattern search range in the face detection process can be narrowed, so that the time required for face detection can be shortened. Will be able to. In such a case, the “non-detection mask” that is the second face detection parameter may not be used. However, if it is desired to further narrow the pattern search range in the face detection process, the “non-detection mask” is used. "Is good to cover the partial image. Furthermore, the partial image may be reduced to a predetermined size as described above, and the reduced image may be used as the detection target image. Alternatively, a predetermined area where the subject's face is considered to exist may be cut out from the reduced image, and the cut-out image may be used as the detection target image.

  In the above description, the “horizontal position” image shown in FIG. 4B is used as the detection target image. For example, as shown in FIG. 8, the “vertical position” image is used as the detection target image. In this case, the following are used for the “non-detection mask” and the “direction pattern”.

  First, the “non-detection mask” is obtained by rotating “non-detection mask 1” to “non-detection mask 3” shown in FIG. 5 by 90 degrees counterclockwise (counterclockwise). The information of each mask used for the “vertical position” image is stored in advance in a storage medium such as a ROM (not shown) of the digital camera.

  Next, when the orientation of the digital camera is “horizontal” as shown in FIG. 8A, “direction pattern 3” shown in FIG. 6B is used as the “direction pattern”. Further, when the posture is “vertical-90 degrees” as shown in FIG. 8B, “direction pattern 1” shown in FIG. 6B is used. Further, when the posture is “vertical +90 degrees” as shown in FIG. 8C, the detection of the face in the direction shown in FIGS. 6A, 6B, 3C, and 2B is enabled. "Direction pattern 4" to be designated as such is newly prepared and used. The information of “direction pattern 4” is stored in advance in a storage medium such as a ROM (not shown) of the digital camera.

  In the above description, the orientations of the digital camera acquired by the orientation information are “horizontal”, “vertical −90 degrees”, and “vertical +90 degrees”, but are not limited to these three types.

  The face detection process described above can also be applied to the case where the subject itself is detected.

(Operational effect of this embodiment)
As described above, in the digital camera of this embodiment, posture information is acquired in the face detection process, and the “non-detection mask” that is the first face detection parameter is determined according to the posture of the camera indicated by the posture information. The Then, the “non-detection mask” determined as the first face detection parameter is placed on the detection target image, and then the face area of the subject is detected from the detection target area of the “non-detection mask”. For this reason, an area in which the face in the detection target image is considered not to exist is designated as a non-detection target area, and face detection processing is not performed for that area.

  Therefore, in the digital camera of the present embodiment, face detection can be performed at a higher speed than when face detection processing is performed on the entire region of the detection target image.

  In the digital camera of the present embodiment, posture information is acquired in the face detection process, and a “direction pattern” that is the second face detection parameter is determined according to the posture of the camera indicated by the posture information. . Then, face detection processing is performed only for faces in the direction in which detection is enabled by the “direction pattern” that is the second face detection parameter, and as a result, the face area of the subject is detected from the detection target image. The For this reason, face detection processing is not performed for a face in a direction that is not designated by the “direction pattern” and that is not likely to be captured.

  Therefore, in the digital camera of the present embodiment, face detection can be performed at a higher speed than when face detection processing is performed on faces in all directions.

  In the digital camera according to the present embodiment, in the face detection process, a reduced image is generated by reducing the image size to a predetermined size based on the image data recorded in the buffer memory 16 by the shooting or image reproduction operation. Is done. Then, the face detection process is performed using the reduced image as the detection target image.

  Therefore, in the digital camera of the present embodiment, the amount of data for pattern search in the face detection process is smaller than when using the original large image size image, so that face detection can be performed at high speed. it can.

  In the digital camera of this embodiment, in the face detection process, a predetermined area where the subject's face is thought to exist is cut out as a partial image from the image data recorded in the buffer memory 16 by shooting or image reproduction. It is. Then, face detection processing is performed using the partial image as a detection target image.

  Therefore, in the digital camera of this embodiment, the pattern search range in the face detection process can be narrowed compared to the case where the image data of the original buffer memory 16 is used as the detection target image. Detection can be performed at high speed.

  In the digital camera of this embodiment, the face area of the subject is detected from the detection target image using face detection parameters determined according to the camera posture indicated by the posture information in the face detection process. The face detection parameters, that is, the first face detection parameter “non-detection mask” and the second face detection parameter “direction pattern” are stored in a storage medium such as a ROM (not shown) of the digital camera. It is stored in advance.

  Therefore, in the digital camera of the present embodiment, face detection can be performed by a simple method that only changes the face detection parameter in accordance with the camera posture indicated by the posture information, and the types of posture supported by the digital camera. Even if there is an addition to the pose, it is possible to cope with the addition of the kind of posture by simply adding the face detection parameter corresponding to it.

  As described above, in the digital camera according to the present embodiment, the object detection process can be speeded up by a simple method.

(Other)
Note that a part or all of the above-described program related to the operation of the digital camera may be executed by an external processing device such as a computer. In that case, a necessary program is installed in the external processing apparatus via a computer-readable storage medium such as a CD-ROM or a communication network such as the Internet.

  In the above description, the embodiment of the digital camera has been described. However, the image playback apparatus of the present invention is also applicable to other devices capable of playing back and displaying an image, such as a mobile phone and a digital video camera. It is possible.

DESCRIPTION OF SYMBOLS 11 ... Imaging lens, 12 ... Lens drive part, 13 ... Imaging device, 14 ... Analog signal processing part, 15 ... Timing generator (TG), 16 ... Buffer memory, 17 ... Image processing part, 18 ... Display control part, 19 ... Display unit, 20 ... control unit, 21 ... sensor unit, 22 ... compression / decoding unit, 23 ... recording interface (recording I / F), 24 ... recording medium, 25 ... operation unit, 26 ... bus

Claims (8)

An imaging device that detects a subject from an image using predetermined subject detection parameters,
Attitude information acquisition means for acquiring information indicating the attitude of the device;
Subject detection means for changing the subject detection parameters in accordance with the posture indicated by the information acquired by the posture information acquisition means when detecting a subject from an image ;
The subject detection unit acquires the information indicating the posture by the posture information acquisition unit, and changes the subject detection parameter in accordance with the acquired posture based on the image. Create a reduced image with the image size reduced to a predetermined size, and detect the subject using the created reduced image
Imaging device comprising a call. The imaging device according to claim 1,
At least one of the subject detection parameters is a parameter that designates a division between a detection target region that is a region where the subject is detected in the image and a detection non-target region that is a region where the subject is not detected. An imaging apparatus characterized by the above. In the imaging device according to claim 1 or 2,
The imaging apparatus according to claim 1, wherein at least one of the subject detection parameters is a parameter for designating a pattern of a subject direction for which detection is effective in the image. In the imaging device according to any one of claims 1 to 3,
The imaging apparatus characterized in that the subject detection means always detects a subject using the image in a lateral position regardless of the posture. A subject detection program for detecting a subject from an image using predetermined subject detection parameters ,
An attitude information acquisition step for acquiring information indicating the attitude of the device;
When detecting a subject from an image, the computer executes a subject detection step of changing the subject detection parameter according to the posture indicated by the information acquired by the posture information acquisition step;
In the subject detection step , information indicating the posture is acquired by the posture information acquisition step, and the subject detection parameter is changed based on the image in parallel with changing the subject detection parameter according to the acquired posture. A subject detection program , wherein a reduced image obtained by reducing an image size to a predetermined size is created , and a subject is detected using the created reduced image. The subject detection program according to claim 5,
At least one of the subject detection parameters is a parameter that designates a division between a detection target region that is a region where the subject is detected in the image and a detection non-target region that is a region where the subject is not detected. An object detection program characterized by Te subject detecting program smell of claim 5 or claim 6,
Is at least one of the parameters before Symbol subject detection, is a parameter that specifies the direction of the pattern of the object to enable detection in the in the image
Subject detecting program characterized and this. In the subject detection program according to any one of claims 5 to 7,
Wherein the object detection step, the subject detecting program, wherein the posture regardless, intends always line detection of an object using the image of the lateral position.

Images