JP2014186634A - Image processing apparatus, image processing method, and image processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately detect minute motions of large and small moving bodies, while preventing false detection.SOLUTION: An image processing apparatus includes: dividing means for dividing first and second image data into a plurality of regions; setting means for acquiring frequency information of the regions divided by the dividing means, and reducing the size of the same region in the first and second image data when both frequency information obtained from the regions are high frequency; changing means which acquires shutter speed and sensitivity at the time of capturing the first and second image data, and increases the size of the region set by the setting means when the acquired shutter speed is lower than a predetermined value and the sensitivity is higher than a predetermined value; and determination means which determines whether motion has occurred in the regions of the first and second image data having the size changed by the changing means.

Description

  The present invention relates to an image processing apparatus, an image processing method, and an image processing program for detecting a local moving area included in images continuously captured by, for example, a digital camera.

  2. Description of the Related Art Image processing apparatuses that detect a subject that moves locally in successive images using a plurality of frame images photographed by a digital camera or the like are known. There is also known an image processing apparatus that performs appropriate image composition by reflecting a detection result of a moving object region in composition processing when combining a plurality of images.

  With regard to a technique for detecting an area including a moving subject, for example, by performing pattern matching on a specific moving body, a means for detecting a position, or a difference value of a plurality of images for each divided block or pixel is obtained. A method of detecting motion using this method is known.

  However, when the target moving body and the background thereof have the same luminance, it is difficult to accurately detect the moving body based on only one value (for example, luminance). In addition, when detecting the movement of a moving object for each divided block, if the unit of division is large, detection of a small moving object cannot be performed accurately, and if the unit of division is small, there is a problem that processing time increases. is there.

  Therefore, means for using a plurality of pieces of information, for example, luminance average, edge density, each color component, and the like, and means for changing the division size for dividing a block for each area have been proposed.

  For example, Patent Document 1 discloses means for calculating a motion vector for each block obtained by dividing two images into a predetermined size, and detecting a moving object region from the vector distribution.

  Further, in Patent Document 2, a block including a moving object is detected from the similarity of each block obtained by dividing a plurality of images into a predetermined size, and the size of the division is changed based on successive detection blocks. Means for detecting are disclosed.

  Furthermore, Patent Document 3 discloses means for changing the block size when encoding image data in units of blocks based on the variance value of the blocks.

JP-A-10-283482 JP 2006-024149 A Japanese Patent Laid-Open No. 09-130801

  However, in Patent Document 1, the motion vector obtained when the moving object disappears or newly enters from the screen becomes uncertain, and as a result, there is a possibility that a correct moving object detection result cannot be obtained.

  Further, in Patent Document 2, since the motion determination is performed with a fine block size when a moving body exists in a continuous block, there is a possibility that the detection accuracy of a small moving body or a moving body with a small motion may be lowered.

  Furthermore, in Patent Document 3, since the block size does not increase in the high frequency region, the detection result of the moving object varies, and a double image appears when applied to the synthesis of a plurality of images, and the image quality deteriorates. there's a possibility that.

  Therefore, an object of the present invention is to provide a mechanism for accurately detecting a minute movement of a moving body while suppressing erroneous detection for a plurality of large and small moving bodies.

  In order to achieve the above object, an image processing apparatus according to the present invention includes a dividing unit that divides each of the first image data and the second image data into a plurality of regions, and each of the regions divided by the dividing unit. Setting means for obtaining the frequency information, and setting the size of the same region to be small when the frequency information obtained from the same region of the first image data and the second image data is both high frequency; The shutter speed and sensitivity at the time of shooting the first image data and the second image data are acquired, and when the acquired shutter speed is slower than a predetermined value and / or the sensitivity is predetermined. If the value is higher than the value, a changing unit that changes the size of the area set by the setting unit to a larger size, and the first after the size is changed by the changing unit Characterized in that it comprises an image data, and a determination unit configured to determine whether there is motion in the region of the second image data.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to detect the small motion of a moving body with high precision, suppressing a misdetection with respect to the large and small moving body.

1 is a block diagram illustrating a schematic configuration example of an image processing apparatus according to a first embodiment of the present invention. FIG. 10 is a flowchart for explaining an operation example of the image processing apparatus. (A) And (b) is a figure which shows an example of two input images. 3A is a diagram illustrating a state in which the image illustrated in FIG. 3A is divided so that the number of blocks is 14 × 10, and FIG. 3B is a diagram illustrating the image illustrated in FIG. It is a figure which shows the state divided | segmented so that it might become a piece. (C) is a figure which shows the state which attached | subjected the number (coordinate) with respect to the divided | segmented block. It is a flowchart figure which shows the process which divides | segments or combines each block based on the frequency information of the block in an even line in the block divided | segmented into 128 * 128 pixels as an initial block size. It is a flowchart figure which shows the process which divides | segments or combines each block based on the frequency information of the block in an odd line in the block divided | segmented into 128 * 128 pixels as an initial block size. It is a figure which shows the example of a change of the block size based on Tv value acquired by the imaging | photography information acquisition part, and ISO sensitivity. It is a figure which shows the example of an image divided | segmented according to the block size of each frequency domain finally determined by step S250 of FIG. (A) is a figure which shows the image which has the area | region determined that the moving body was included by the motion determination part, (b) is a figure which shows the image from which the conifer was detected, (c) is the detection in a motion determination process It is a figure which shows a result. (A) And (b) is a figure which shows an example of two input images in the image processing apparatus which is the 2nd Embodiment of this invention, (c) and (d) are the images shown to (a), respectively, FIG. 5B is a diagram illustrating a result of detecting a face area in the image illustrated in (b). It is a figure which shows the example of a change of the block size based on Tv value acquired by the imaging | photography information acquisition part, and ISO sensitivity. It is a figure which shows the example of an image divided | segmented according to the block size of each frequency domain finally determined by step S250 of FIG. It is a figure which shows the image which has the area | region determined by the motion determination part to include the mobile body. It is a figure which shows the image synthesize | combined by the image synthetic | combination part.

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

(First embodiment)
FIG. 1 is a block diagram showing a schematic configuration example of an image processing apparatus according to the first embodiment of the present invention.

  As shown in FIG. 1, the image processing apparatus according to the present embodiment includes an image input unit 100, a subject detection unit 101, a block size determination unit 102, a frequency analysis unit 103, a shooting information acquisition unit 104, an image division unit 110, and a motion determination. Unit 130 and image composition unit 140.

  In the image input unit 100, images continuously captured by an imaging device such as a digital camera are sequentially input. The subject detection unit 101 detects the position of a specific subject from the input image. The block size determination unit 102 determines a block size when the input image is divided for each region. The frequency analysis unit 103 acquires frequency information of the input image.

  The shooting information acquisition unit 104 acquires information at the time of shooting such as ISO sensitivity, aperture value, and Tv value from the input image. The image dividing unit 110 divides an input image into blocks of a predetermined size in the H (horizontal) and V (vertical) directions. The motion determination unit 130 determines whether there is an area including a moving object for each block divided by the image division unit 110. The image composition unit 140 composes a plurality of images and generates a composite image.

  Next, an operation example of the image processing apparatus having the above configuration will be described with reference to FIG. Each process in FIG. 2 is executed by the CPU by loading an image processing program stored in a ROM, a hard disk, or the like of the image processing apparatus into the RAM.

  In step S200, when a plurality of images are input from the image input unit 100, the CPU proceeds to step S210. Here, it is assumed that two images 300 and 310 (see FIG. 3) successively captured by an imaging device such as a digital camera are sequentially input. Here, the image 300 corresponds to an example of the first image data of the present invention, and the image 310 corresponds to an example of the second image data of the present invention.

  In step S <b> 210, the CPU determines whether a human face is detected in the two images 300 and 310 by the subject detection unit 101. Here, assuming that the face of a person has not been detected, the process proceeds to step S220.

  In step S220, the CPU determines the block size for the two images 300 and 310 by the block size determination unit 102, and proceeds to step S230. Specifically, the CPU determines the block size by the block size determination unit 102 so that the input image is divided into 14 in the H (horizontal) direction and 10 in the V (vertical) direction as a predetermined block size. .

  In step S230, the CPU divides the images 300 and 310 with the block size determined in step S220 by the image dividing unit 110, and proceeds to step S240.

  The image 301 shown in FIG. 4A and the image 311 shown in FIG. 4B have the number of blocks of 14 × 10 from the image 300 shown in FIG. 3A and the image 310 shown in FIG. It is divided so that it becomes a piece. Here, one block size is 128 × 128 pixels. An image 320 shown in FIG. 4C is obtained by assigning numbers (coordinates) to the divided blocks.

  In step S240, the CPU uses the frequency analysis unit 103 to obtain frequency information using a known technique such as a bandpass filter that extracts a specific frequency band for each block divided in step S230. Then, the CPU sets a block size based on the acquired frequency information, and proceeds to step S250.

  In step S240, specifically, based on the frequency information acquired by the frequency analysis unit 103, each block is classified into three types of bands: high frequency, medium frequency, and low frequency. Here, when the frequency information obtained from the same block area of the two images 301 and 311 is different, the result classified as lower frequency is used. This is because the frequency information of the subject, not the background, is reflected in the block size for the area including the moving object. Further, according to the classified frequency band, the block size is determined as high frequency: 64 × 64 pixels, medium frequency: 128 × 128 pixels, and low frequency: 256 × 256 pixels.

  With reference to FIG.5 and FIG.6, the process of step S240 is further explained in full detail. FIG. 5 is a flowchart showing a process of dividing or combining the blocks based on the frequency information of the blocks on the even lines in the block divided into 128 × 128 pixels as the initial block size.

  In FIG. 5, in step S <b> 241, the CPU determines whether the frequency information of the target block is a high frequency, medium frequency, or low frequency region. The CPU proceeds to step S242 if the frequency information of the block of interest is determined to be a high frequency region, and proceeds to step S243 if it is determined to be a low frequency region. On the other hand, if the CPU determines that the frequency information of the block of interest is the middle frequency region, the CPU ends the process without dividing or combining the blocks.

  In step S242, the CPU divides the block determined to be in the high frequency region into x1 / 2 in the H (horizontal) and V (vertical) directions, and ends the process.

  In step S243, the CPU determines whether or not the blocks positioned on the right, the lower, and the lower right are in the low frequency region with respect to the block determined to be in the low frequency region. If the CPU determines that the right, lower, and lower right blocks are in the low frequency area, the CPU proceeds to step S244. If the CPU determines that the blocks are not in the low frequency area, the CPU divides or combines the blocks. The process is finished without.

  In step S244, the CPU combines the four blocks determined as the low frequency region in step S243 to form new blocks, and ends the process.

  After the above processing is performed for all blocks of even lines, the blocks are divided or combined based on the frequency information of the blocks on the odd lines. FIG. 6 is a flowchart showing a process of dividing or combining the blocks based on the frequency information of the blocks on the odd lines in the block divided into 128 × 128 pixels as the initial block size.

  In FIG. 6, in step S245, the CPU determines whether the frequency information of the block of interest is a high frequency, medium frequency, or low frequency region. If the CPU determines that the frequency information of the block of interest is a high frequency region, the CPU proceeds to step S246. If the CPU determines that the frequency information is a low frequency region, the CPU proceeds to step S247. On the other hand, if the CPU determines that the frequency information of the block of interest is the middle frequency region, the CPU ends the process without dividing or combining the blocks.

  In step S246, the CPU divides the block determined to be in the high frequency region into x1 / 2 in the H (horizontal) and V (vertical) directions, and ends the process.

  In step S247, the CPU determines whether or not the blocks located on the right, the lower, and the lower right are in the low frequency region and are uncombined with respect to the block determined to be in the low frequency region. If the CPU determines that the blocks located on the right, lower, and lower right are in the low frequency region and are not combined, the CPU proceeds to step S248 and determines that the block is in the low frequency region and is not unconnected. In this case, the process is terminated without dividing or combining the blocks.

  In step S248, the CPU combines the four blocks determined to be uncombined in the low frequency region in step S247 to form new blocks, and ends the process.

  By performing the above processing for all blocks of odd lines, the block size for all regions is set. As a result, in the motion determination process described later, the sensitivity can be increased even with a small amount of motion as the frequency is higher, and it is possible to easily detect minute motions of the moving body. .

  Returning to FIG. 2, in step S <b> 250, the CPU acquires the Tv value (shutter speed) and ISO sensitivity at the time of shooting the input images 300 and 310 using the shooting information acquisition unit 104. Then, the CPU changes to the block size determined in step S240 based on the acquired Tv value and ISO sensitivity, and proceeds to step S260.

  In step S250, specifically, when the Tv value of the captured image is 1/125 seconds or more which is a predetermined value with respect to the block size set in step S240, that is, when the shutter speed is slow, H , V direction is doubled respectively. If the Tv value of the photographed image is smaller than the predetermined value of 1/640 seconds with respect to the block size determined in step S240, the H and V directions are each halved.

  This is because the amount of motion is small when the shooting interval between the two images 300 and 310 is short, so that the motion can be controlled by increasing the sensitivity to even a small amount of motion during the motion determination process described later. This is to facilitate detection. However, the block size is not changed here, assuming that the Tv value acquired by the imaging information acquisition unit 104 is 1/250.

  If the sensitivity of the captured image is ISO 800 or more with respect to the block size determined in step S240, the H and V directions are each doubled. This is to suppress erroneous determination due to the influence of noise at the time of high sensitivity during the motion determination process described later.

  FIG. 7 shows an example of changing the block size based on the Tv value acquired by the imaging information acquisition unit 104 and the ISO sensitivity. FIG. 8 is a diagram illustrating an example of an image divided according to the block size of each frequency region finally determined in step S250. As shown in FIG. 8, a high frequency area such as a leaf is divided by a small block size of 64 × 64 pixels, and a low frequency area such as a background sky is divided by a large block size of 256 × 256 pixels. I understand that.

  In step S260, the CPU determines whether or not there is movement for each of the divided blocks with respect to the input images 300 and 310 by the movement determination unit 130, and proceeds to step S270.

  In step S260, specifically, the motion determination unit 130 calculates the absolute value of the difference between the luminance component and the color difference component for each pixel with respect to the block images in the same region in the images 301 and 311, and further, Find the sum of. Then, the motion determination unit 130 compares the sum of absolute differences of the luminance component and color difference component for each block determined in step S220 with a threshold, and if at least one of the sum of absolute differences is equal to or greater than the threshold, It is determined that the block is an area of the moving body. That is, if any one of the following equations (1) to (3) is satisfied, it is determined that the region is a moving object.

  In the above formulas (1) to (3), Img1 and Img2 are the luminance component (Y) and color difference components (U, V) of the pixel values of the input first image 300 and second image 310, respectively. The difference absolute value is calculated for all the pixels in the block, and the sum is obtained. Th_Y, Th_U, and Th_V are threshold values for luminance components and color difference components for determination.

  FIG. 9A is a diagram illustrating an image 600 having regions 600 a to 600 c that are determined by the motion determination unit 130 to include a moving object. As shown in FIG. 9 (a), it can be seen that the sum of the absolute values of the differences is large for the region where the car and the person are moving, and is determined to be a moving body region.

  In step S270, the CPU performs a composition process on the images 300 and 310 by the image composition unit 140 using the motion determination result for each block obtained in step S260, and ends the process.

  In the synthesis process here, the image synthesis unit 140 outputs the average value of the pixels of the images 300 and 310 as a new pixel value for the area that is not determined to be included by the motion determination unit 130. To do. As a result, a composite image with reduced noise can be obtained. The image composition unit 140 outputs the pixel value of the image 300 as it is without performing composition for the region determined by the motion determination unit 130 to include the moving object. This makes it possible to prevent a double image from appearing due to the synthesis.

  In this embodiment, the case where the input image is divided so that the number of blocks is 14 × 10 is exemplified, but the number of blocks may be larger or smaller than 14 × 10. In this embodiment, all blocks are square blocks, but may be vertically or horizontally long blocks depending on the distribution of the high frequency region.

  Furthermore, in the present embodiment, the case where the average value of the pixels of the images 300 and 310 is output when the image is synthesized with respect to the area that is not determined to include the moving object is illustrated. Alternatively, the weighted addition may be performed based on the weighting coefficient calculated from the sum of the above.

  In addition, when a unit for detecting a specific high-frequency region such as a coniferous tree from an image is provided, it may be determined that the region is moving with respect to the entire region when the motion determination result in the region varies.

  That is, the frequency and color information for each region in the input image is analyzed, and when the value of the high frequency signal and the sensitivity signal is large, it is determined that the tree is a conifer and includes means for outputting the position and size of the region And FIG. 9B is a diagram illustrating an image 610 in which a conifer is detected. When the detection result 620 shown in FIG. 9C is obtained by the movement determination process in step S260 due to the movement of the leaves, there is a possibility that the switching of the two input images is noticeable and unnatural at the time of synthesis.

  Therefore, when there is a block determined to be moving even in the coniferous region, it is determined that the entire region is moving as in the determination result 630 illustrated in FIG. It is possible to suppress such unnaturalness.

  As described above, in the present embodiment, when the moving object region is detected in units of blocks, the block size to be divided is changed according to the frequency information of the subject and the shooting information such as Tv value and sensitivity. As a result, it is possible to accurately detect a minute movement of the moving body while suppressing erroneous detection for a plurality of large and small moving bodies. Further, even when a plurality of images are combined, it is possible to generate a combined image without the moving subject being doubled.

(Second Embodiment)
Next, an image processing apparatus according to the second embodiment of the present invention will be described with reference to FIGS. In the present embodiment, detailed description of the same parts as those in the first embodiment will be omitted while diverting the drawings and symbols of the first embodiment.

  In the image processing apparatus according to the present embodiment, a case where a human face is detected from two images by the subject detection unit 101 in step S210 of FIG. 2 is taken as an example.

  In step S200, when an image is input from the image input unit 100, the CPU proceeds to step S210. Here, it is assumed that two images 700 and 710 (see FIG. 10A and FIG. 10B) successively captured by an imaging device such as a digital camera are sequentially input.

  In step S210, the CPU determines whether a human face is detected in the two images 700 and 710 by the subject detection unit 101 using a known face image recognition circuit or the like. Here, it is assumed that a human face has been detected, and the process proceeds to step S220. In the image 701 shown in FIG. 10C and the image 711 shown in FIG. 10D, the face areas are detected with respect to the image 700 shown in FIG. 10A and the image 710 shown in FIG. 10B, respectively. The results are shown.

  Here, when a face is detected from the image, the subject detection unit 101 outputs the detected face and eye coordinate positions and sizes. At this time, it is assumed that 768 × 640 pixels are obtained as the size of the face region.

  In step S220, the CPU determines the block size for the two images 700 and 710 by the block size determination unit 102, and proceeds to step S230. In step S220, the block size determination unit 102 determines the position and size of the eyes based on the output of the subject detection unit 101 for the eye region, and the output of the subject detection unit 101 for the mouth region. To estimate the position and size of the mouth. Then, the block size determination unit 102 sets 256 × 128 pixels as a block size that includes each of the eyes and mouth that are expected to move. Thereby, blinking of eyes and movement of the mouth are detected as one area.

  The block size determination unit 102 sets a small block size of 64 × 64 pixels for the area other than the eyes and mouth in the face area. This is to make it easier to detect a motion by increasing the sensitivity to a slight amount of motion of a person.

  In step S230, the CPU divides the images 700 and 710 by a predetermined block size by the image dividing unit 110, and proceeds to step S240. Here, the image dividing unit 110 divides the face area with the block size determined in step S220, and for the area other than the face area, 128 × 128 as in the first embodiment. Divide by pixel block size.

  In step S240, the CPU uses the frequency analysis unit 103 to obtain frequency information using a known technique such as a bandpass filter that extracts a specific frequency band for each block divided in step S230. Then, similarly to the first embodiment, the CPU determines a block size based on the acquired frequency information, and proceeds to step S250.

  In step S250, the CPU acquires the Tv value and ISO sensitivity at the time of shooting of the input images 700 and 710 by the shooting information acquisition unit 104. Then, the CPU changes to the block size determined in step S240 based on the acquired Tv value and ISO sensitivity, and proceeds to step S260.

  In the present embodiment, since the Tv value is 1/30, the block size determined in step S240 is doubled in the H and V directions. However, for face regions and facial organ regions such as eyes and mouth, the block size is not changed according to the Tv value, and the block size is doubled in the H and V directions only at high sensitivity (ISO 800 or higher). . This is because, in portrait photography and the like, a person's face often has a small amount of motion, and for the motion detection in the face area, a slight motion can be detected by using the block size determined in step S240. It is for doing so.

  FIG. 11 shows an example of changing the block size based on the Tv value acquired by the photographing information acquisition unit 104 and the ISO sensitivity. FIG. 12 is a diagram illustrating an example of an image divided according to the block size of each frequency region finally determined in step S250. As shown in FIG. 12, the face area of a person is divided by a block size corresponding to the size of each organ regardless of imaging information, and a fixed size of 64 × 64 pixels is used in the area other than each organ. You can see that it is divided by.

  In step S260, the CPU uses the motion determination unit 130 to determine whether or not there has been a motion for each divided block with respect to the input images 700 and 710, as in the first embodiment, and step S270. Proceed to

  FIG. 13 is a diagram illustrating an image 800 having regions 800a and 800b determined by the motion determination unit 130 to include moving objects. As shown in FIG. 13, the sum of the absolute values of the difference between the two images 700 and 710 is large and the motion is determined in the area where the eye blinks and the peripheral area of the slightly moved face. I understand that.

  In step S270, using the motion determination result for each block obtained in step S260, the CPU performs a composition process on the images 700 and 710 by the image composition unit 140 in the same manner as in the first embodiment. To end the process.

  Here, when a person blinks during the two shooting periods, the image composition unit 140 outputs the pixel value of the image 700 as it is without performing the composition process on the eye region, so that the image is obtained by composition. It becomes possible to prevent becoming unnatural. FIG. 14 is a diagram illustrating an image synthesized by the image synthesis unit 140. As shown in FIG. 14, it can be seen that two images 700 and 710 can be synthesized without being double-synthesized with respect to the area of the moving object.

  In the present embodiment, the case where a human face is detected as the characteristic part of the subject is illustrated, but the characteristic part to be detected is not limited to the face. Further, if there is at least one block including the face area among the blocks determined as the moving body area, all the blocks including the face area may be determined as the moving body area. By using one input image for the entire face area at the time of image synthesis, it is possible to obtain a more natural synthesized image without being affected by movement.

  In addition, when a means for displaying the input image on the electronic finder at a predetermined frame rate and tracking the moving body by a known technique is provided, the block size is determined based on the moving direction and speed of the obtained moving body. You may decide. By applying an appropriate block size to the area including the moving object, it is possible to detect motion with high accuracy.

  As described above, in the present embodiment, when the moving body region is detected in units of blocks, the block size to be divided is changed according to the size of the characteristic part of the detected subject. As a result, it is possible to accurately detect a minute movement of the moving body while suppressing erroneous detection for a plurality of large and small moving bodies. Further, even when a plurality of images are combined, it is possible to generate a combined image without the moving subject being doubled.

  In addition, this invention is not limited to what was illustrated by said each embodiment, In the range which does not deviate from the summary of this invention, it can change suitably.

  The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed. It can also be realized by executing software (program) acquired via a network or various storage media on a personal computer (CPU, processor).

DESCRIPTION OF SYMBOLS 100 Image input part 101 Subject detection part 102 Block size determination part 103 Frequency analysis part 104 Shooting information acquisition part 110 Image division part 130 Motion determination part 140 Image composition part

Claims (6)

Dividing means for dividing the first image data and the second image data into a plurality of regions,
While acquiring the frequency information of each area | region divided | segmented by the said division means, when all the frequency information obtained from the same area | region of said 1st image data and said 2nd image data is high frequency, Setting means for setting the size small;
The shutter speed and sensitivity at the time of shooting of the first image data and the second image data are acquired, and when the acquired shutter speed is slower than a predetermined value, and / or the sensitivity is predetermined. A change means for changing the size of the area set by the setting means to a larger size,
Image processing comprising: determination means for determining whether or not each area of the first image data and the second image data after the size is changed by the changing means has moved. apparatus. Based on a determination result by the determination unit, comprising a combining unit that combines the first image data and the second image data;
The synthesizing unit newly sets an average value of pixels of the first image data and the second image data for an area that has not been determined to have moved by the determining unit during the synthesizing process. The pixel value of the first image data is output as it is without being synthesized for an area which is output as a pixel value and determined to have moved by the determination means. The image processing apparatus described. Detecting means for detecting a characteristic part of a subject and a part where movement of the characteristic part is expected from the first image data and the second image data;
The dividing unit divides the feature part detected by the detection unit so that a size of a region including a part that is expected to move is larger than a size of a region other than the region of the feature part. The image processing apparatus according to claim 1 or 2.   The determination means determines that all areas of the feature portion have moved when there is at least one region determined to have moved in the feature portion detected by the detection means. The image processing apparatus according to claim 3. A dividing step of dividing each of the first image data and the second image data into a plurality of regions;
While acquiring the frequency information of each area | region divided | segmented by the said division | segmentation step, when all the frequency information obtained from the same area | region of said 1st image data and said 2nd image data is high frequency, A setting step to set the size small,
The shutter speed and sensitivity at the time of shooting of the first image data and the second image data are acquired, and when the acquired shutter speed is slower than a predetermined value, and / or the sensitivity is predetermined. A change step for changing the size of the area set in the setting step to a larger size,
A determination step for determining whether or not there is a movement in each area of the first image data and the second image data after the size is changed in the changing step. Method. A dividing step of dividing each of the first image data and the second image data into a plurality of regions;
While acquiring the frequency information of each area | region divided | segmented by the said division | segmentation step, when all the frequency information obtained from the same area | region of said 1st image data and said 2nd image data is high frequency, A setting step to set the size small,
The shutter speed and sensitivity at the time of shooting of the first image data and the second image data are acquired, and when the acquired shutter speed is slower than a predetermined value, and / or the sensitivity is predetermined. A change step for changing the size of the area set in the setting step to a larger size,
Determining whether or not there is a movement in each area of the first image data and the second image data after the size is changed in the changing step; Image processing program.