JP2017049662A - Information processing apparatus, control method thereof, program, and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve accuracy of processing for acquiring position information of an end of a predetermined operating body, from three-dimensional position information obtained on the operating body.SOLUTION: An information processing apparatus acquires three-dimensional position information on an operating body which exists in a space on a predetermined operation surface, detects a first point corresponding to an end of the operating body, on the basis of the acquired three-dimensional position information of the operating body, detects a second point corresponding to the end of the operator, on the basis of a shape of the operating body captured in an area of interest including the position of the first point, in a two-dimensional image obtained by capturing the space, and recognizes two-dimensional position information of the first point in parallel to the operation surface, or two-dimensional position information of the second point in parallel to the operation surface, as a position on the operation surface pointed by the operating body.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a technique for recognizing a touch operation based on a proximity state between a recognition target and a touch target surface.

  Device operation based on gesture recognition technology that operates a UI (user interface) in accordance with the movement and position of a human hand detected using various cameras and sensors is becoming widespread. Even in a table top interface in which an image or UI is displayed on a table surface and the image or UI is operated by touching it with a hand or a pen, the end of a predetermined operation body such as a fingertip or a pen tip is not a touch panel. The method of detecting the state of touching is beginning to be used.

  In Patent Document 1, in a process of extracting a pixel group having a skin color component from a two-dimensional data acquired by a camera as a hand region in which a hand appears as a subject, a three-dimensional image acquired by a three-dimensional scanner when extraction fails. Use data to detect and complement the hand shape.

JP 2007-241833 A

  When using 3D image data to obtain 3D position information of a subject as in Patent Document 1, conversion to coordinates is performed assuming that each pixel value of the image data corresponds to position information in the depth direction. Do. When recognizing a subject with a certain width in the depth direction or a plurality of overlapping subjects, a contour is extracted from the data image by setting a threshold value for the coordinate value in the depth direction. can do. However, when the resolution in the depth direction is low, or when an error occurs in the conversion from pixel information to position information due to factors such as the environment, it is difficult to accurately obtain the contour of the subject. When using 3D image data to detect the degree of proximity to the table surface, such as the fingertip or pen tip, the edge of the operation object can be regarded as touching the table surface, If they are different, it is easy to cause misrecognition. For example, the position specified by the touch operation is shifted from the actual fingertip position.

  The present invention has been made in view of the above, and improves the accuracy of processing for obtaining position information of an end recognized as an operation position from three-dimensional position information obtained with respect to a predetermined operation tool. Objective.

  In order to solve the above-described problems, an information processing apparatus according to the present invention images an operation object acquisition unit that acquires three-dimensional position information regarding an operation object existing in a space on a predetermined operation surface, and the space. Image acquisition means for acquiring a two-dimensional image, and first detection means for detecting a first point corresponding to an end of the operation body based on the three-dimensional position information acquired by the operation body acquisition means And setting means for setting a region of interest including the position of the first point detected by the first detection unit in the two-dimensional image, and imaging the region of interest by the setting unit of the two-dimensional image. A second detection means for detecting a second point corresponding to the end of the operating body based on the shape of the operating body, and the operation of the first point detected by the first detecting means. Two-dimensional position information parallel to the surface Recognizing means for recognizing one of the two-dimensional position information parallel to the operation surface of the second point detected by the second detection device as a position on the operation surface indicated by the operation body. And comprising.

  According to the present invention, the accuracy of processing for obtaining position information of an end recognized as an operation position from three-dimensional position information obtained with respect to a predetermined operation body is improved.

The figure showing an example of the table top interface system using information processor 100 The figure showing an example of hardware constitutions and functional composition of information processor 100 The flowchart which shows an example of the flow of the main process which the information processing apparatus 100 in 1st Embodiment performs The flowchart which shows an example of the flow of the determination process of the operation position in 1st Embodiment. The figure showing the example of the state of a user's hand, and the example of the hand area detected from a distance image corresponding to it The figure showing the example of the state of a user's hand, and the example of the hand area detected from a distance image corresponding to it The figure which shows an example of the determination process of an operation position The figure which shows an example of the determination process of an operation position The figure which shows an example of the determination process of an operation position The figure which shows an example of the determination process of an operation position The flowchart which shows an example of the flow of the determination process of the operation position in the modification 2. The figure which shows an example of the determination process of an operation position The flowchart which shows an example of the flow of the main process of the information processing apparatus in 2nd Embodiment. The flowchart which shows an example of the flow of a hand region determination process The figure which shows an example of the determination process of a hand region

  Hereinafter, information processing according to an embodiment of the present invention will be described in detail with reference to the drawings. In addition, the structure described in an Example is an illustration and is not the meaning which limits the scope of the present invention to those structures.

<First Embodiment>
First, as a first embodiment, an example of processing for recognizing a touch operation performed by an operator on an item projected on a table surface of a table top interface system will be described.

  FIG. 1A is an example of the appearance of a tabletop interface system in which the information processing apparatus 100 according to the present embodiment is installed. The operation surface 101 is a table portion of the table top interface, and the operator can input a touch operation by touching the operation surface 101. However, in this embodiment, a touch sensor is not mounted on the operation surface 101. For this reason, the information processing apparatus 100 does not determine whether the operation surface 101 and an operation body such as a finger or a pen operated by the user are actually in contact with each other, but a state close enough to be regarded as being in contact (hereinafter referred to as touch input). The touch operation can be recognized by detecting the touch state.

  In the present embodiment, the distance image sensor 102 is installed above the operation surface 101 so as to look down at the operation surface. A distance image is an image in which information corresponding to the distance from the reference position (for example, the center of a lens, etc.) of the imaging means that captures the distance image to the subject surface imaged by the pixel is reflected in the value of each pixel. It is. In the present embodiment, the distance image sensor 102 captures the distance value from the distance image sensor 102 to the operation surface 101 or the object surface above the distance image. The captured distance image is input to the information processing apparatus 100 as a distance image. The information processing apparatus 100 acquires the three-dimensional position of the operator's hand 106 by analyzing the distance image, and recognizes the input operation. Therefore, the operator moves the space gesture by moving a predetermined object such as a hand in a range that can be imaged by the distance image sensor 102 in the space on the operation surface (the space between the operation surface 101 and the distance image sensor 102). An operation can be entered. In the present embodiment, a sensor that acquires distance information using a reflection pattern (or reflection time) of infrared light is used. However, for example, a distance image can be obtained by installing a stereo camera system or an infrared light emitting element and an infrared light receiving element. In addition, in the space including the operation surface 101, any means capable of obtaining three-dimensional position information including the height direction of the operation body is not limited to a mode in which a distance image is captured. The present embodiment can also be implemented by a method for obtaining original position information.

  In the present embodiment, the visible light camera 103 is installed so as to look down on the operation surface 101 from above. The information processing apparatus 100 can function as a document camera that controls the visible light camera 103 to capture an image of an object placed on the operation surface 101 and obtain a read image thereof. The information processing apparatus 100 detects and further identifies an object existing in the space on the operation surface 101 based on a visible light image obtained by the visible light camera 103 and a distance image obtained by the distance image sensor 102. The object includes, for example, an operator's hand, a paper medium, a document such as a book, and other three-dimensional objects. However, in the case of the system illustrated in FIG. 1A, the angle of view of the distance image sensor 102 and the visible light camera 103 does not include the operator's head around the table. Therefore, in the obtained distance image, the end of the image intersects with some part of the user's arm (portion beyond the shoulder).

  The projector 104 projects an image on the upper surface of the operation surface 101. In this system, the operator performs an operation by touch or space gesture on the item 105 included in the projected image. As described above, in the present embodiment, the distance image acquired by the distance image sensor 102 is used for detecting the hand 106 and recognizing the operation. By using the distance image, there is an advantage that even if the color of the operator's hand changes due to the projection light of the projector 104, the distance image is hardly affected. The display device of the present system can be configured by using the operation surface 101 as a liquid crystal display instead of the projector 104. In that case, even if a method of detecting a human hand from an image by detecting a skin color region from a visible light image, the hand can be detected without being affected by the projection light.

  If the image obtained by viewing the operation surface 101 from above is obtained, the distance image sensor 102 and the visible light camera 103 themselves do not necessarily have to be installed above. For example, reflected light is imaged using a mirror. You may comprise so that it may do. Similarly, in the example of FIG. 1A, the projector 104 projects onto the operation surface 101 so as to look down from diagonally above, but the projection light projected in different directions is used using a mirror or the like. It may be reflected on the operation surface 101. The same applies to the case where the operation surface 101 is installed along the vertical direction.

  In the present embodiment, the x, y, and z axes shown in FIG. 1 are defined in a three-dimensional space on the operation surface 101, and position information is handled. In the example of FIG. 1A, the point 107 is the origin of the coordinate axes. Here, as an example, the two dimensions parallel to the upper surface of the table are the xy plane, and the direction perpendicular to the table upper surface and extending upward is the positive direction of the z axis. In the present embodiment, the z-axis direction corresponds to the height direction in the world coordinate system. However, the present embodiment is also applied to a system in which a non-horizontal surface such as a whiteboard or a wall surface is used as the operation surface 101, a case where the operation surface has irregularities, or a virtual surface generated using MR. Is possible.

  FIG. 1B is a diagram illustrating the relationship between the distance image captured by the distance image sensor 102 and the system. In the present embodiment, the distance image sensor 102 has a function of acquiring a distance image reflecting three-dimensional position information of a predetermined space included in the angle of view and a two-dimensional infrared image different from the distance image. . In the present embodiment, the distance image sensor 102 emits infrared light from the distance image sensor 102 toward the operation surface 101, and the reflected light reflected by the surface of the subject is received by the light receiving element of the sensor.

  The distance image used in this embodiment reflects the distance to the subject surface corresponding to the time taken to reflect each pixel by measuring the phase delay of the received reflected light in the pixel value. Is. In the present embodiment, since the distance image sensor 102 captures a distance image so as to look down the operation surface 101, the distance information represented by the pixel value of the distance image is in the depth direction in the direction of looking down the operation surface 101 from the distance image sensor 102. Distance. In other words, each pixel value of the distance image includes information for obtaining position information (z coordinate) in the height direction from the operation surface 101. However, in the present embodiment, the distance image sensor 102 is installed so as to have an angle with respect to the z axis. Therefore, in this embodiment, the pixel value of the distance image is not used as it is as the z-coordinate, but is defined in the image by performing coordinate conversion using parameters according to the sensor and the installation environment as described later. The position information is used after being converted into xyz coordinates of world coordinates. The distance calculation method used by the distance image sensor 102 may be an infrared pattern projection method or a parallax method.

  The two-dimensional infrared image used in the present embodiment has the intensity of reflected light received by the light receiving element of the sensor as a pixel value. A similar infrared image can also be obtained by receiving the reflected light reflected by the subject surface by the infrared light included in the ambient light without irradiating the infrared light from the distance image sensor 102. Further, the two-dimensional infrared image may be other than the infrared image, and a color image or a gray scale image captured by the visible light camera 103 may be acquired.

  In the present embodiment, the two-dimensional infrared image and the distance image obtained from the distance image sensor 102 have the same size (for example, 640 [dot] × 480 [dot]), and all the pixels correspond to each other. . The reflection time (distance information) or intensity of infrared light reflected from the same position of the same subject is reflected in the pixels whose coordinates on the plane match. Note that if a distance image and a two-dimensional infrared image having the above relationship are obtained, it is not necessary for the image capturing means of the both to be integrated as in the distance image sensor 102 of the present embodiment. That is, a sensor that obtains three-dimensional position information including a distance and an image sensor that captures a two-dimensional image may be provided separately.

  In FIG. 1B, an image 108 represents an example of the content of a distance image captured by the distance image sensor 102. However, here, the distance information reflected in the pixel value is omitted, and only the edge of the subject is specified. The operation surface 101 and the origin 107 correspond to FIG. A range 111 corresponds to an angle of view of the distance image sensor 102. Hereinafter, the image edge 111 is referred to.

  As shown in FIG. 1B, a two-dimensional coordinate system based on the u axis and the v axis is set in the distance image. In the example of FIG. 1B, the resolution of the distance image is 640 [dot] × 480 [dot]. Assume that the position coordinates in the distance image of the operation position 110 are (u, v), and the pixel value corresponding to the distance from the distance image sensor 102 to the operation position is d. In the present embodiment, coordinate conversion based on the lens characteristics of the distance image sensor 102 and the relative positional relationship with the operation surface 101 is performed on the position information defined in the distance image in this way. Thereby, the coordinate of each pixel is mapped to the coordinate system in the real world defined in the table, and the three-dimensional position (x, y, z) in the real space can be acquired for the operation position 110. The transformation matrix used for coordinate transformation is acquired by performing adjustment work in advance when the distance image sensor 102 is installed. Note that the distance d used for calculating the three-dimensional coordinates is not limited to a single pixel value, but a noise removal process and an averaging process are performed on pixels corresponding to several pixels near the operation position 110 in the hand region. And may be specified.

  A hatched area 109 is an image of the operator's hand 106 in the distance image 108 (hereinafter simply referred to as the hand area 109). In the present embodiment, by performing threshold processing on the z-coordinate, an area in which a subject existing at a position higher than the operation surface 101 that is the table surface is captured as a hand area. In this embodiment, one operation position is detected for each detected hand. The operation position is a coordinate of a position estimated that the operator is pointing with a finger. In the present embodiment, as a premise, when the operator designates a point for a touch operation, it is prescribed to take a “pointing pose” with one finger extended. This is because a pose with one finger extended is a natural posture for many people to point to one point. Therefore, as the operation position, a position estimated to be an end portion of the hand region 109 is specified. In a pointing pose, the edge hits the fingertip. Specifically, the hand region 109 is extracted from the distance image, and the coordinate indicating the pixel located farthest from the image end 111 in the hand region 109 is regarded as one point corresponding to the fingertip. In the case of the hand region 109 shown in FIG. 1B, the point farthest from the image end 111 is specified as the operation position 110. Note that the method for specifying the operation position is not limited to this, and for example, the five fingers of the hand may be detected from the hand region to specify the end of the predetermined finger.

  In the present embodiment, the center of the portion where the hand region 109 intersects the image end 111 in each frame of the distance image repeatedly captured according to the frame rate is defined as the hand entry position. In the case of the hand region 109 in FIG. 1B, the intrusion position 112 is specified. The intrusion position 112 is also acquired as three-dimensional position information in real space by converting the position information defined in the distance image. In the present embodiment, when a plurality of hand regions are detected from the distance image, the operation position and the intrusion position are specified for each of them. Then, the touch operation is recognized for each of the plurality of hand regions. However, when the fingertip exists at a height close to the operation surface 101, the difference between the pixel values of the table and the fingertip is extremely small. Therefore, depending on the size of the resolution and the detection error, it is difficult to accurately distinguish between the finger and the table by the above threshold processing. For example, if the threshold is set higher than the actual table surface and the tip of the finger is considered to be a portion below the threshold, the specified fingertip position is not actually the fingertip, It can be the root. When this detection result is used as an operation position as it is, the operation position that the user intends to instruct and the detected operation position are deviated. Therefore, an item different from the item that the user wants to select may be in a selected state, which may cause a decrease in operability.

  Therefore, in the present embodiment, not only the three-dimensional position information obtained from the distance image but also the position information of the operation position in the operation surface direction is detected using a separately acquired two-dimensional image. More specifically, in order to identify position information in a direction parallel to the operation surface among the three-dimensional position information of the operation position, from both the distance image and the two-dimensional infrared image, the end of the operation body To detect the point and its position information. Then, which position is recognized as the position instructed by the operation is selected according to the situation.

  Hereinafter, in this specification, it is assumed that the operator's hand 106 and its finger are used as an example of an operation body and an end portion thereof used by the operator for input of a touch operation. However, in the present embodiment, the present invention can be applied not only to fingers but also instruments such as a stylus and a robot arm as the operating body. Note that the end of the operating body refers to a part used to point to the operation position for the touch operation. However, if the part belongs to a part of the operating body and can be used for input of the touch operation, the end of the operating body has a protruding shape. It does not need to be limited to the end.

  FIG. 2A is a hardware configuration diagram of a tabletop interface including the information processing apparatus 100 according to the present embodiment. The central processing unit (CPU) 200 is connected to the system bus 204 by using the RAM 202 as a work memory, executing an OS and a program stored in the ROM 201 and the storage device 203, performing various processing calculations and logical determinations, and the like. Control each configuration. The storage device 203 is a hard disk drive, an external storage device connected by various interfaces, and the like, and stores programs and various data related to the operation recognition processing of the embodiment. The distance image sensor 102, under the control of the CPU 200, captures a distance image of a space on the operation surface 101 including a table on which the item is displayed and an operator's hand operating the item, and the captured distance image is a system bus 204. Output to. In the present embodiment, the distance image acquisition method will be described based on a reflection time method (Time-of-Flight method) that is less affected by ambient light or table surface display. It is also possible to use a pattern method or the like. The projector 104 projects and displays an image item to be operated on the table according to the control of the CPU 200.

In the above-described system, the visible light camera 103, the distance image sensor 102, and the projector 104 are external devices connected to the information processing apparatus 100 via an input / output interface, respectively. Configure the processing system. However, these devices may be integrated into the information processing apparatus 100. FIG. 2B is an example of a block diagram illustrating a software configuration of the information processing apparatus 100. Each of these functional units is realized by the CPU 200 developing a program stored in the ROM 201 in the RAM 202 and executing processing according to each flowchart described later. The execution result of each process is held in the RAM 202. Further, for example, when hardware is configured as an alternative to software processing using the CPU 200, arithmetic units and circuits corresponding to the processing of each functional unit described here may be configured.

  The three-dimensional information acquisition unit 210 acquires a distance image captured by the distance image sensor 102 at regular intervals according to the frame rate, and stores the acquired distance image in the RAM 202 as needed. The target acquired by the three-dimensional information acquisition unit 210 and exchanged with each functional unit is actually a signal corresponding to the image data, but will be described as simply “acquiring a distance image” in the present specification.

  The operating tool acquisition unit 211 performs threshold determination and noise reduction processing on each pixel of the distance image acquired by the three-dimensional information acquisition unit 210, and extracts a hand region in the distance image. The hand region is a pixel group in which a hand used as an operating body by an operator is captured as a subject in the input distance image. The first detection unit 212 identifies one point corresponding to the end of the operation tool as the first point based on the contour information of the hand region extracted by the operation tool acquisition unit 211, acquires a coordinate value, and stores the coordinate value in the RAM 202. Hold. For the first point, three-dimensional coordinate information is obtained based on the distance image. In the present embodiment, coordinates indicating a pixel point that is farthest from the image end 111 in the contour of the hand region are detected as the first point. At this time, in the first detection unit 212 of the present embodiment, the center of the portion where the hand region and the image end 111 intersect is used as the intrusion position of the hand region.

  The two-dimensional image acquisition unit 213 acquires a two-dimensional infrared image captured by the distance image sensor at regular intervals according to the frame rate, and holds the input image in the RAM 202 as needed. In practice, a process corresponding to a signal corresponding to image data is performed, but in the present specification, it is simply described as “acquiring a two-dimensional infrared image”. In the present embodiment, the two-dimensional infrared image used as the input image is captured by the same imaging unit as the distance image, but they are different images. The region setting unit 214 sets a region of interest including a pixel corresponding to the first point detected by the first detection unit 212 in the input image. In the present embodiment, a vector corresponding to the direction of the fingertip is generated, and a region of interest including the tip where the vector extends is set. Based on the shape of the subject in the region included in the region of interest set by the region setting unit 214 in the input image, the second detection unit 215 sets one point corresponding to the end of the operating body as the second point. The coordinate value is acquired and stored in the RAM 202. However, the coordinates obtained from the two-dimensional infrared image are two-dimensional. In the present embodiment, when the input state of the touch operation is detected based on the height (z coordinate) of the first point, it is instructed using the two-dimensional coordinates (x, y coordinates) of the second point. The process of recognizing existing position information is performed. Therefore, height information is not necessarily required for the second point. However, when it is necessary to use position information in the three-dimensional direction, the pixel value of the pixel corresponding to the second point in the distance image is corrected as described with reference to FIG. The vertical coordinate (z coordinate) may be calculated.

  The recognition unit 216 recognizes an operation input by the end of the operation body. In the present embodiment, the position of the end of the operating body is referred to as the operation position. The operation position is determined based on the position information of the first point or the second point. In the present embodiment, the first point is preferentially determined as the operation position. Then, according to the degree of proximity of the operation position detected from the distance image to the operation surface 101, it is determined whether the information processing apparatus 100 is in a touch input state by the operation tool. For example, when the distance (z coordinate corresponding to the height) between the first point and the operation surface 101 is smaller than a predetermined threshold, it is determined that touch input is being performed. Then, by tracking the position on the operation surface 101 instructed by the operation body while the touch input is being performed, a touch event such as touch, release, move, flick, pinch, and rotate is generated, and the display control unit 217 receives the touch event. Notice. However, as the position on the operation surface 101 designated by the operation body, either one of the two-dimensional position information (xy coordinates) parallel to the operation surface of the first point or the second point is selected and used. To do. Furthermore, the recognition unit 216 determines the shape of the operation tool region. For example, the number of fingers, finger direction, hand direction, and the like are determined. The display control unit 217 generates drawing data reflecting the result of each process executed in response to a user operation notified by the recognition unit 216 and outputs the drawing data to the projector 104 to control the content displayed on the operation surface. To do. The user operation is defined by the operation position, the movement of the operation position, the shape (pose) of the operation body region, and the like, and can input some instruction to the information processing apparatus 100.

  FIG. 3 is a flowchart showing the flow of main processing of the information processing apparatus according to this embodiment. In this embodiment, when the information processing apparatus 100 is turned on, first, in step S301 of FIG. 3, the display control unit 217 generates a screen on which UI components are arranged based on the data acquired from the storage device 203. And output to the projector 104. Then, the projector 104 projects a screen on the upper surface of the operation surface 101.

  Next, in step S <b> 302, the three-dimensional information acquisition unit 210 acquires a distance image from the distance image sensor 102. Further, the two-dimensional infrared image and the distance image acquired by the two-dimensional infrared image acquisition unit 213 are acquired. Incidentally, the distance image sensor 102 captures a two-dimensional infrared image and a distance image of the operation area on the upper surface of the operation surface 101 at predetermined time intervals while the information processing apparatus is powered on.

  In step S303, the operating tool acquisition unit 211 detects a predetermined operating tool (a user's hand or arm). When the operating tool is detected (S303, YES), the process proceeds to step S304. On the other hand, when the operating tool is not detected (S303, NO), the process returns to step S302, and the processes of steps S302 and S303 are repeated until the operating tool is detected.

  In step S304, an operation position determination process is executed. The contents of step S304 will be described later with reference to the flowchart of FIG.

  In step S305, the recognition unit 216 recognizes the operation input to the information processing apparatus 100 based on the operation position determined in step S304. For example, when the height (z coordinate) representing the degree of proximity of the operation position to the operation surface is below a predetermined threshold, the touch operation is recognized. Then, the information indicating the recognition result is notified to the display control unit 217. At this time, the position (xy coordinates) on the operation surface indicated by the operation body by touching is the xy coordinates of the operation position determined in step S304. In step S306, the display control unit 217 updates the display of objects (images, data, UI, etc.) on the screen projected by the projector 104 in accordance with the recognized operation. In step S307, it is determined whether or not the information processing apparatus has been turned off. When the power is turned off (S306, YES), the information processing apparatus 100 ends all the processes. On the other hand, if the power is not turned off (S306, NO), the process returns to step S302.

  FIG. 4 is a flowchart showing the operation position determination process (S304) in the present embodiment. First, in step S401, the first detection unit 212 detects the first point as the operation position based on the hand region acquired by the operating tool acquisition unit 211. In the present embodiment, the center of the intersection of the hand area and the image edge 111 is set as the intrusion position, and the pixel farthest from the intrusion position among the pixels included in the hand area is detected as the first point regarded as the fingertip position in the distance image. To do. Here, the detected position information of the first point is three-dimensional position information (x, y, z coordinates) obtained based on the distance image. Accordingly, in step S401, three-dimensional position information (x, y, z coordinates) is obtained as the position information of the operation position.

  In step S402, the recognition unit 216 compares the height (z coordinate) of the operation position acquired in step S401 with a threshold T1 representing a predetermined height, and determines whether it is equal to or less than the threshold T1. Here, the threshold value T1 is a threshold value for determining the validity that the first point is obtained by accurately detecting the position of the fingertip. As described above, when the fingertip is approaching the operation surface 101, that is, when the height of the fingertip from the operation surface 101 is low, the possibility that the contour of the hand region is not accurate due to the size of the resolution of the sensor and the detection error is compared. Increase. Therefore, in step S402, it is determined whether or not it is necessary to perform processing for improving the accuracy of the fingertip position using a means different from the distance image. In the present embodiment, the case where the height of the operation position is equal to or less than the threshold value T1 is regarded as the case where it is necessary to perform processing for improving the accuracy of the fingertip position using means different from the distance image. When the height of the operation position is equal to or less than the threshold T1 (S402, YES), the position information of the operation position is further accurately determined using the two-dimensional infrared image in the processing after step S403. If the height of the operation position is greater than the threshold value T1 (S402, NO), the process of FIG. 4 is terminated and the process proceeds to S305.

  Next, in step S403, the second detection unit 215 acquires information indicating the direction of the fingertip based on the distance image. The fingertip direction is a direction starting from the center position of the hand and ending at the operation position. The center position of the hand is the center point of the widest area of the hand area. In the present embodiment, the fingertip direction is obtained as a vector. In step S <b> 404, the region setting unit 214 determines a region including the fingertip direction vector as the attention region from the operation position in the two-dimensional infrared image. In step S405, the second detection unit 215 detects the second point within the attention area of the two-dimensional infrared image. Specifically, edge extraction is performed within the attention area, and an area surrounded by the boundary between the edge and the attention area and including the operation position is detected as a fingertip area. Let the edge part of the detected fingertip area | region be the 2nd point which can be considered that it corresponds to a fingertip. The position information detected here is two-dimensional position information (xy coordinates) parallel to the operation surface 101. Since edges may be interrupted, it is preferable to perform a complementing process that connects between the disconnected edges. However, the method of detecting the fingertip region may be background difference using color information or moving region detection.

  In step S406, the recognition unit 216 determines whether the second point satisfies a predetermined condition. Here, the predetermined condition is that two-dimensional coordinate information parallel to the operation surface 101 is valid among the three-dimensional position information of the operation position defined based on the first point in step S401, or This is a condition for determining whether correction is to be made with the position information of the second point. For example, when the first point is detected from the distance image and the contour shape of the portion close to the operation surface 101 of the hand region is not accurately obtained, the value obtained as the xy coordinates of the first point is The actual fingertip may be off. On the other hand, the xy coordinates of the second point are values obtained from a two-dimensional infrared image different from the distance image, and it is unlikely that the above problem has occurred. In step S406, it is determined which position of the first point or the second point should be adopted as the position designated on the operation surface by the operation surface. In the present embodiment, as the predetermined condition, it is assumed that the predetermined condition is satisfied when the distance from the fingertip direction vector of the second point is smaller than the reference value. This determination result means that it is appropriate to correct the second point by the fingertip position method. On the other hand, when the predetermined condition is not satisfied, it means that the second point cannot be said to be more appropriate because the second point and the first point are largely separated. When the distance from the fingertip direction vector of the second point is close and the predetermined condition is satisfied (S406, YES), the process proceeds to S407. On the other hand, when the distance from the fingertip direction vector of the second point is far and the predetermined condition is not satisfied (S406, NO), the processing of the flowchart of FIG.

  In step S407, the first detection unit 212 corrects the xy coordinates in the three-dimensional position information of the operation position with the xy coordinates of the second point at the position of the second point detected in S404. Then, the process of the flowchart of FIG. 4 ends. As described above, in the present embodiment, when the process of step S407 is executed, when the operation recognition process of step S305 is executed, the xy coordinate of the second point is the operation surface 101 indicated by the fingertip. Recognized as the upper position. On the other hand, when the process of step S407 is not executed, when the operation recognition process of step S305 is executed, the xy coordinate of the first point is recognized as the position on the operation surface 101 indicated by the fingertip. Is done.

  Here, when the threshold value T1 described above is set as a value smaller than the threshold value T2 for determining whether the fingertip is considered to be touching the operation surface 101, the z-coordinate of the operation position is set in step S407. The z coordinate detected for the first point may be maintained after processing. This is because step S407 is executed only when the z coordinate of the first point is equal to or less than the threshold value T1, and therefore it is obvious that the z coordinate falls below the threshold value T2. However, not only for the purpose of comparison with the touch threshold value T2, but also when the z-coordinate needs to be obtained for the second point, the pixel value is referenced with reference to the pixel corresponding to the xy-coordinate of the second point in the distance image. The z coordinate may be calculated. In this case, either the three-dimensional position information of the first point or the three-dimensional position information of the second point is selectively recognized as the operation position.

  Next, operations performed by the information processing apparatus 100 according to the present embodiment will be described with reference to FIGS. 5 to 8 in association with specific examples of the user's hand state. FIGS. 5 to 8 explain step by step how the user touches the operation surface 101 and inputs a touch operation after reaching out over the operation surface 101. The number of the step in which characteristic processing is performed in each stage is shown in parentheses.

  First, a state where the user's finger is at a position away from the operation surface 101 will be described with reference to FIGS. FIG. 5A shows a part of the distance image at time t1, and FIG. 5B is a view of the space on the operation surface 101 at time t1 as seen from the side surface (y direction). First, a distance image and a two-dimensional infrared image at time t1 are acquired from the distance image sensor 102 (S302). The hand area detected from the acquired distance image is shown as a gray area in FIG. Here, the operating tool at time t1 is referred to as operating tool A. The operating tool A is a user's hand and arm. The hand is a part of the arm that extends beyond the wrist (finger side) and includes parts such as five fingers, palm, and back.

  First, in the present embodiment, the first point is detected as the operation position from the hand region detected from the distance image (S401). As shown in FIG. 5A, first, an intrusion position p1 is detected, and a pixel farthest from the intrusion position p1 in the hand region is detected as a first point p2. Further, the three-dimensional position (x2, y2, z2) in the real space of the first point p2 is acquired based on the position information and the pixel value in the distance image. Then, it is determined whether the height (z2) from the operation surface of the first point p2 is equal to or less than the threshold value T1 (S402). In the case of FIG. The operation is recognized using the obtained three-dimensional position (x2, y2, z2) as the operation position. In the present embodiment, in order to recognize the touch operation, the height (z coordinate) of the operation position that represents the degree of proximity between the fingertip and the operation surface 101 is compared with the threshold value T2. When the z coordinate is less than the threshold T2, it is determined that the degree of proximity between the fingertip and the operation surface 101 is close to the degree that the touch state can be considered. In the state of FIG. 5A, the operation tool A is maintained at a position sufficiently higher than the operation surface 101, and the height z2 of the operation tool is larger than the threshold value T2, and thus is not regarded as a touch state.

  Next, an operation example of the information processing apparatus 100 in a state in which the user's fingertip is in contact with the operation surface 101 will be described with reference to FIGS. However, here, a case where the accuracy of the periphery of the fingertip among the contours of the hand region obtained from the distance image is lowered due to the fingertip being very close to the operation surface 101 will be particularly described. FIG. 6A shows a part of the distance image acquired at time t2. FIG. 6B is a diagram of the space on the operation surface 101 viewed from the y direction at time t2. Time t2 is a time after time t1.

  A distance image and a two-dimensional infrared image at time t2 are acquired from the distance image sensor 102 (S302). The hand region acquired from the acquired distance image is shown as a gray region in FIG. The operating tool at time t2 is referred to as operating tool B. In the hand region at this time, a part corresponding to a part of the fingertip is missing in the operating body B. This is because the resolution or detection accuracy of the distance image sensor 102 is not sufficient, and it becomes difficult to extract the boundary between the overlapped operation surface 101 and the user's finger based on the distance information. Also in this case, in this embodiment, the intrusion position p3 is detected based on the hand area, and the pixel farthest from the intrusion position p3 included in the hand area is detected as the first point p4 (S304). Further, the three-dimensional position (x4, y4, z4) in the real space of the detected first point p4 is acquired as the three-dimensional position information of the operation position.

  Here, it is assumed that the z coordinate corresponding to the height from the operation surface 101 among the coordinates (x4, y4, z4) of the detected first point p4 is smaller than the threshold T1. At this time, in this embodiment, when the height of the operation position from the operation surface is determined to be equal to or less than the threshold T1 (YES in S402), the two-dimensional position information parallel to the operation surface of the operation position is converted into the two-dimensional information. It will also detect from an infrared image. Specifically, first, a fingertip direction vector A of the operating tool B is created in the distance image (S403). FIG. 7A is a diagram illustrating an outline of a method for determining a fingertip direction in a distance image, and FIG. 7B is a diagram illustrating an overview of a method for determining a region of interest in a two-dimensional infrared image. As shown in FIG. 7A, in the present embodiment, the center position of the widest place of the operating tool B is acquired as the hand center position p5. Then, a fingertip direction vector A having the acquired center position p5 of the hand as a start point and the first point p4 as an end point is created. Next, as shown in FIG. 7B, a part of the two-dimensional infrared image is set as a region of interest (S404). In this embodiment, the starting point of the fingertip direction vector A calculated from the distance image is changed to the first point p4 detected from the distance image, and the fingertip direction vector B is created. Instead of simply extending the fingertip direction vector A, the fingertip search range can be narrowed down by considering the realistic finger length range by moving the starting point. However, since the distance between the first point and the center position of the hand changes depending on how the finger is bent, in some cases, the fingertip direction vector A is extended to a predetermined length without moving the starting point. A direction vector B may be set.

  In the present embodiment, the partial area including the fingertip direction vector B is determined as the attention area on the two-dimensional infrared image. In the case of the present embodiment, the distance image and the two-dimensional infrared image have a one-to-one correspondence with the pixels. Therefore, the fingertip direction vector B defined in the distance image can be copied to the two-dimensional infrared image as it is. it can. FIG. 8A is an enlarged view of the attention area determined in the two-dimensional infrared image. In the present embodiment, the second point is detected based on the contour shape of the subject appearing in the attention area (S405). Specifically, edge extraction is performed in the attention area, and an area that is surrounded by a boundary line between the edge and the attention area and that includes the first point p4 detected in the distance image is defined as a fingertip area. A pixel serving as an end portion of the detected fingertip region is specified as a second point p6. Then, position information (x6, y6) in two dimensions parallel to the operation surface 101 of p6 is acquired. If necessary, p6 in the distance image is referred to, and three-dimensional position information (x6, y6, z6) of p6 including the z coordinate is acquired.

  Based on the detected coordinates of the second point p6, it is determined at this time whether the coordinate information of the operation position defined by the coordinate information of the first point should be corrected (S406). In the present embodiment, it is determined as a predetermined condition that the distance d1 between the second point p6 and the fingertip direction vector B in the xy plane is equal to or less than the threshold T3, and correction is executed when this condition is satisfied. In the case of FIG. 8A, since the distance d1 is equal to or less than the threshold value T3, (x6, y6) of the second point p6 is regarded as valid as the xy coordinates of the operation position (YES in step S406). . Then, the operation point is corrected at the first point p4 and the second point p6 (step S407). If necessary, the z-coordinate z6 of the second point may be obtained from the distance image.

  FIG. 8B shows the corrected operation position and its three-dimensional coordinates. However, FIG. 8B shows a case where the z coordinate z6 of the corrected operation position is obtained using the pixel value of the distance image indicated by the xy coordinate of the second point. As described above, when the height z4 of the first point p4 is lower than the threshold value T1, the height z6 of the second point z4 is lower than the touch threshold value T2. Therefore, regardless of whether or not z6 has been obtained, in step S305, the xy coordinate of the second point is recognized as the position on the operation surface 101 designated by the user's fingertip. For example, among the UI buttons projected on the operation surface 101, the button including the designated position (x6, y6) is selected, and the projection content is updated accordingly (step S306).

  As described above, in the present embodiment, the position information of the end of the operating tool detected based on the three-dimensional position information is supplemented with the position information of the end of the operating tool detected from the two-dimensional image according to the situation. At this time, the range of searching for the position of the end of the operating tool in the two-dimensional image is limited to the inside of the attention area determined by the position information of the end of the operating tool detected based on the three-dimensional position information. Thereby, the total calculation amount can be reduced and the first point detection accuracy can be improved.

<Modification 1>
Here, as a first modification of the first embodiment, an example in which a region of interest is set by another method will be described with reference to FIGS. 9 and 10. In FIGS. 9 and 10, the same numbers are assigned to elements that overlap those in FIGS. 7 to 8.

  First, as shown in FIG. 9A, an area having a predetermined size centered on the first point p4 detected from the distance image is set as an attention area. Here, the attention area is set to an area that is large enough to encompass the fingertip, wherever the detected first point is actually from the end of the finger to the root. As the size of the attention area, a predetermined value is selected according to the height (z4) of the first point p4. For example, the size of the region of interest may take into account the distance between the distance image sensor 102 and the hand (first point or the center position of the hand) and the size of the hand region. However, the degree of reflection of the arm varies greatly depending on the frame. Therefore, when considering the size of the hand area, define an area that represents the size of the hand (wrist to tip) part of the arm that is unlikely to differ in size between frames, and handle the size. . For example, the width of the hand region near the center position of the hand is set as the size of the hand region. In addition, if the hand area is large to some extent (corresponding to the case where the hand is close to the sensor), the attention area is set to be large. Select the size of the region of interest, such as setting.

  Next, FIG. 9B is an enlarged view of the region of interest. The center-of-gravity position p7 (x7, y7, z7) of the operating tool B within the attention area is calculated. A fingertip direction vector C having the calculated center of gravity position p7 as a start point and the first point p4 as an end point is created. Next, as shown in FIG. 10A, the starting point of the created fingertip direction vector C is changed to the first point p4, and the fingertip direction vector D is created. Then, a partial area including the fingertip direction vector D in the two-dimensional infrared image is set as the attention area.

  FIG. 10B is an enlarged view of the region of interest in the two-dimensional infrared image set according to the first modification. In the first embodiment and the modification, even if the second point can accurately extract the actual fingertip position, if the distance from the fingertip direction vector is long, the second point is set as the operation position. Is not used. Therefore, it is desirable that the fingertip direction vector represents the direction of the finger more accurately. However, since the fingertip direction vector B is based on the center position of the hand, the difference between the actual finger direction and the finger may be large depending on conditions such as the finger being operated and the shape of the hand at the time of operation. For example, comparing FIG. 8A and FIG. 10B, the fingertip direction vector D of the first modification is closer to the actual fingertip direction than the fingertip direction vector B of the first embodiment. I understand that. On the other hand, the calculation amount required for the process of generating the vector is smaller in the fingertip direction vector B method than in the fingertip direction vector D. The attention area setting method may be selectively used depending on whether the information processing apparatus 100 or the application should prioritize the processing speed.

<Modification 2>
Furthermore, as a second modification, a method for determining a region of interest without using a fingertip direction vector will be described. FIG. 11 is a flowchart showing an operation position determination process (S304) in the second modification. The same processing steps as those in the flowchart of FIG. 4 are given the same numbers, and detailed description thereof is omitted.

  In the second modification, when the first point detected from the distance image in step S401 is detected and the height is larger than the threshold value T1, the process proceeds to step S1100. In step S1100, the region setting unit 214 determines a region of interest having a predetermined size on the basis of the operation position at this time point specified from the xy coordinates of the first point in the two-dimensional infrared image. . FIG. 12A shows a state in which a region of interest C having a predetermined size is determined in the two-dimensional infrared image with the position of the first point p4 detected in the distance image as the center. Here, the size of the attention area is set to a larger attention area when the hand area is somewhat large (corresponding to the case where the hand is close to the sensor) as in the idea of setting the attention area in the first modification. Is small (corresponding to the case where the sensor is far from the hand), the attention area is set small.

  In step S <b> 1101, the second detection unit 215 detects a hand region B, which is a pixel group in which the operating tool appears as a subject, from a portion in the attention region C of the two-dimensional infrared image. Specifically, edge detection is performed in the attention area C, and an area surrounded by the boundary line between the detected edge and the attention area C and including the first point p4 is acquired as the hand area B.

  In step S1102, the second detection unit 215 similarly sets the attention area C having a predetermined size centered on the first point in the distance image, and the hand area detected from the distance image and the attention area. An area included in C is a hand area A (FIG. 12A). It is determined whether the difference in shape between the hand region A and the hand region B acquired in step S1101 can be regarded as small. Specifically, the area of the hand region A and the hand region B that do not overlap is calculated as the area D, and the determination is performed by comparing the area D with the predetermined threshold value D1. For example, when the area D is equal to or smaller than the threshold value D1, the difference in shape between the hand area A and the hand area B is considered to be small (S1102, YES), and the process proceeds to step S1103. When the area D is larger than the threshold value D1, it is considered that the size difference between the hand region A and the hand region B is large (NO in S1102), and the process returns to the flowchart of FIG. 3 without correcting the operation position.

  In step S1103, the second detection unit 215 sets, as the second point p6, the position regarded as the end of the operating body based on the contour shape of the hand region B in the attention region C of the two-dimensional infrared image. To detect. In step S1104, the recognition unit 216 corrects the xy coordinates of the operation position with the xy coordinates of the second point p6. And it returns to the flowchart of FIG. According to the second modification, the first point can be corrected by the shape of the hand region without detecting the fingertip direction vector.

<Second Embodiment>
The information processing apparatus according to the first embodiment identifies an operation position and recognizes an operation based on a position specified on the operation surface. In the second embodiment, an operation corresponding to the posture of the user's hand, which is the operation body, is recognized based on the shape of the operation body (hand region). For example, by identifying the pointing pose described above, the posture in which the fingers of one hand are stretched and the two are bent, the posture in which all the fingers are stretched, and the like, it is associated with each posture in advance. Recognize as an instruction. Further, for example, various instructions are recognized based on the shape of the movement locus of the operating body in a state where each posture is maintained.

  The second embodiment can be executed by the information processing apparatus 100 having the same configuration as that of the second embodiment. However, the main process is performed according to the flowchart of FIG. 13 instead of the flowchart of FIG. Only the differences from FIG. 3 will be described here.

  In the flowchart of FIG. 13, in step S1301, a process for correcting the hand area detected in step S303 is performed. FIG. 14 is a flowchart showing hand region determination processing (S1301) in the second embodiment. First, in step S1401, the second distance acquisition unit 206 acquires the height z8 of the lowest point p8 (first point) of the hand region C, and when the height z8 is determined to be equal to or less than the threshold T1 (S1401). , YES), the process proceeds to S1402.

  In step S1402, the region setting unit 214 determines the attention region D with reference to the lowest point p8 of the hand region C. As shown in FIG. 15A, an attention area D having an arbitrary size is determined around the lowest point p8 detected in the distance image. The size of the attention area is preferably larger as the distance is shorter and smaller as the distance is farther in consideration of the distance to the distance image sensor.

  In step S1403, the second operating tool detection unit 310 detects the hand region D from the determined two-dimensional infrared image in the attention region D. Specifically, edge detection is performed in the attention area D, and an area surrounded by the boundary line between the detected edge and the attention area D and including the lowest point p8 is detected as the hand area D. In step S1404, the second operating tool detection unit 310 sets the hand region D in the attention region D of the two-dimensional infrared image as the hand region.

  Returning to the flowchart of FIG. 13, in step S <b> 1302, the recognition unit 216 determines from the shape of the hand region D which posture is registered in the dictionary. Then, an instruction associated with the determined posture is recognized. For example, as shown in FIG. 15B, when it is determined that the shape of the hand region D is a posture in which three fingers are extended, it is recognized as an instruction to display a list of operation menus. In that case, in step S306, the display control unit 217 executes display update for projecting the operation menu onto the operation surface 101. According to the present embodiment, it is possible to accurately recognize not only an operation for instructing the coordinates of a point as an operation position but also an operation input according to the posture of the operating body.

<Other embodiments>
The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in a computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

210 Three-dimensional information acquisition unit 211 Operation object acquisition unit 212 First detection unit 213 Two-dimensional image acquisition unit 214 Region setting unit 215 Second detection unit 216 Recognition unit 217 Display control unit

Claims (17)

An operating tool acquisition means for acquiring three-dimensional position information relating to an operating tool existing in a space on a predetermined operating surface;
Image acquisition means for acquiring a two-dimensional image obtained by imaging the space;
First detecting means for detecting a first point corresponding to an end of the operating body based on the three-dimensional position information acquired by the operating body acquiring means;
Setting means for setting a region of interest including the position of the first point detected by the first detection means in the two-dimensional image;
Second detection means for detecting a second point corresponding to the end of the operating body based on the shape of the operating body imaged in the region of interest by the setting means in the two-dimensional image;
Two-dimensional position information parallel to the operation surface of the first point detected by the first detection means and two-dimensional position information parallel to the operation surface of the second point detected by the second detection means. Recognizing means for recognizing any of position information as a position on the operation surface instructed by the operation body;
An information processing apparatus comprising:   The operating tool acquisition means is a means for acquiring a pixel group constituting an area in which the operating tool appears as a subject from a distance image in which distance information in the space is reflected in each pixel value, and the distance is The information processing apparatus according to claim 1, wherein the distance corresponds to a degree of proximity to an operation surface included in the space.   The recognizing means is detected by the first detecting means when the position of the first point detected by the first detecting means is closer than a predetermined distance to the operation surface existing in the space. One of the two-dimensional position information parallel to the operation surface of the first point and the two-dimensional position information parallel to the operation surface of the second point detected by the second detection means. The information processing apparatus according to claim 1, wherein the information processing apparatus is recognized as a position on the operation surface instructed by the operation body. If the position of the first point detected by the first detecting unit is not considered to be closer than a predetermined distance to the operation surface existing in the space, the recognizing unit detects the first point Recognizing as a position on the operation surface instructed by the operating body based on the position information of the point,
Two-dimensional position information parallel to the operation surface of the first point and the operation of the second point when the operation surface existing in the space is considered to be closer than a predetermined distance. The process of determining whether to recognize any of two-dimensional position information parallel to a surface as a position on the operation surface instructed by the operation body is performed. Information processing device.   The recognition means recognizes a state in which the end of the operating body is closer to the operation surface than a predetermined distance as a state in which the end of the operating body is touching the operation surface. The information processing apparatus according to any one of claims 1 to 4.   The recognizing means recognizes a touch operation input to the information processing device based on a position while the end of the operating body is in a state of touching the operation surface. The information processing apparatus according to claim 5. The operating body is a user's arm that inputs an operation to the information processing apparatus, and the first point corresponding to the end is a point corresponding to a fingertip of the user's arm. The information processing apparatus according to claim 1, wherein the information processing apparatus is an information processing apparatus. The recognizing unit is configured to calculate one of the three-dimensional position information of the first point detected by the first detecting unit and the three-dimensional position information of the second point detected by the second detecting unit. The information processing apparatus according to claim 1, wherein the information processing apparatus is recognized as three-dimensional position information of an operation position corresponding to an end of the predetermined operation body. The recognition means includes a contour shape of a portion corresponding to the attention area set in the two-dimensional image among areas where the operation body acquired by the operation body acquisition means from the distance image is captured as a subject, When the difference between the contour shape of the operating body imaged in the attention area in the two-dimensional image is larger than a predetermined reference, two-dimensional position information parallel to the operating surface of the first point, the operating body If the difference is smaller than a predetermined reference, two-dimensional position information of the first point parallel to the operation surface and the operation of the second point are recognized. The process of determining whether any of two-dimensional position information parallel to a surface is recognized as a position on the operation surface instructed by the operation body is performed. Information processing device. The two-dimensional image is an image obtained by imaging infrared light,
The said recognition means acquires the shape of the outline of the said operation body by detecting an edge inside the said attention area among the said two-dimensional images, The any one of Claim 1 thru | or 9 characterized by the above-mentioned. The information processing apparatus described. An operation object acquisition step of acquiring three-dimensional position information related to the operation object existing in the space on the predetermined operation surface by the operation object acquisition means;
An image acquisition step of acquiring a two-dimensional image obtained by imaging the space by an image acquisition means;
A first detection step of detecting a first point corresponding to an end of the operation body based on the three-dimensional position information acquired in the operation body acquisition step by a first detection unit;
A setting step of setting a region of interest including the position of the first point detected in the first detection step in the two-dimensional image by the setting means;
Second detection means detects a second point corresponding to the end of the operating body based on the shape of the operating body imaged in the region of interest in the setting step in the two-dimensional image. Two detection steps;
Two-dimensional position information parallel to the operation surface of the first point detected in the first detection step by the recognition means, and the operation surface of the second point detected in the second detection step. Recognizing any of the two-dimensional position information parallel to the position as the position on the operation surface instructed by the operation body,
A method for controlling an information processing apparatus, comprising: Computer
An operating tool acquisition means for acquiring three-dimensional position information relating to an operating tool existing in a space on a predetermined operating surface;
Image acquisition means for acquiring a two-dimensional image obtained by imaging the space;
First detecting means for detecting a first point corresponding to an end of the operating body based on the three-dimensional position information acquired by the operating body acquiring means;
Setting means for setting a region of interest including the position of the first point detected by the first detection means in the two-dimensional image;
Second detection means for detecting a second point corresponding to the end of the operating body based on the shape of the operating body imaged in the region of interest by the setting means in the two-dimensional image;
Two-dimensional position information parallel to the operation surface of the first point detected by the first detection means and two-dimensional position information parallel to the operation surface of the second point detected by the second detection means. Recognizing means for recognizing any of position information as a position on the operation surface instructed by the operation body;
A program for causing an information processing apparatus to function.   A computer-readable storage medium storing the program according to claim 12. First image acquisition means for acquiring a distance image reflecting three-dimensional position information defined in a space on a predetermined operation surface on which a person performs an operation using a hand;
Second image acquisition means for acquiring a two-dimensional image obtained by imaging the space;
First detection means for detecting a hand region in which the human hand appears as a subject from the distance image based on the three-dimensional position information reflected in the distance image;
In the two-dimensional image, a setting unit that sets a region of interest so as to include a reference position among the hand regions acquired by the first detection unit;
Second detection means for detecting a hand region in which the human hand appears as a subject from the attention region set by the setting means in the two-dimensional image;
Recognition means for recognizing the state of the person's hand based on either the shape of the hand area acquired by the first detection means or the shape of the hand area acquired by the second detection means;
An information processing apparatus comprising: A first image acquisition step of acquiring a distance image in which three-dimensional position information defined in a space on a predetermined operation surface on which a person performs an operation using a hand is reflected by a first image acquisition unit;
A second image acquisition step of acquiring a two-dimensional image obtained by imaging the space by a second image acquisition means;
A first detection step of detecting a hand region in which the human hand appears as a subject from the distance image based on the three-dimensional position information reflected in the distance image by a first detection means;
A setting step of setting a region of interest so as to include a reference position among the hand regions acquired in the first detection step in the two-dimensional image by the setting unit;
A second detection step of detecting a hand region in which the human hand appears as a subject from the region of interest set in the setting step by the second detection means;
A recognition step of recognizing the state of the person's hand based on either the shape of the hand region acquired in the first detection step or the shape of the hand region acquired in the second detection step. When,
A method for controlling an information processing apparatus, comprising: Computer
First image acquisition means for acquiring a distance image reflecting three-dimensional position information defined in a space on a predetermined operation surface on which a person performs an operation using a hand;
Second image acquisition means for acquiring a two-dimensional image obtained by imaging the space;
First detection means for detecting a hand region in which the human hand appears as a subject from the distance image based on the three-dimensional position information reflected in the distance image;
In the two-dimensional image, setting means for setting a region of interest so as to include a reference position among the hand regions acquired by the first detection unit;
Second detection means for detecting a hand region in which the human hand appears as a subject from the attention region set by the setting means in the two-dimensional image;
Recognition means for recognizing the state of the person's hand based on either the shape of the hand area acquired by the first detection means or the shape of the hand area acquired by the second detection means;
A program for causing an information processing apparatus to function.   A computer-readable storage medium storing the program according to claim 16.

Images