JP2014192838A - Ar gesture user interface system for portable terminal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a user interface system capable of performing interlocking to control of a portable terminal or operation of an application within the terminal while feeling an actual touch of a virtual object by displaying the virtual object while overlapping with a real image photographed by a compact camera mounted on the back of the portable terminal, and operating the virtual object with a finger of the hand placed in a space in the depth of the terminal.SOLUTION: A space in the depth of a terminal T where a hand (h) is placed is photographed by a compact camera 1 mounted on the back of the terminal T, motions of a finger (f) are recognized by processing an obtained image sequence, and a virtual keyboard (k) and a recognition result are superimposed on the image sequence and displayed on a display 2 of the terminal T. Thus, an operation region becomes a wide space in the depth of the terminal T and terminal operation with higher convenience can be implemented.

Description

  The present invention superimposes and displays a virtual object on a real image captured by a small camera attached to the back of the mobile terminal, and operates it with a finger placed in a space behind the terminal, The present invention relates to a user interface system that can be linked to control and operation of an application in a terminal.

  In recent years, computers have become smaller and lighter, and high-performance portable terminals such as smartphones have become popular. As a result, operations such as document creation and image editing, which have been conventionally performed on a PC, can be easily performed outdoors regardless of location. However, since a small portable terminal has a small operation area installed on the surface thereof, there is a limit to the operation with a finger for work that requires a cumbersome operation, and convenience is lowered.

  As a means to solve this problem, the augmented reality (AR) technology is used to move the two-dimensional operation area on the surface of the small device into a three-dimensional real space, and a wider operation area. Research has been conducted to improve the convenience of equipment by ensuring the above. In Non-Patent Documents 2, 3, and 5, virtual objects are projected onto the surface of an object in real space, a user's hand or arm, and the like can be operated in real space. There are problems such as the need to be worn on the body and limitations on the shape of the projection surface of the virtual object.

  Non-Patent Document 1 proposes a system that superimposes a virtual object on a real image captured by a camera and obtains a feeling of directly touching the virtual object with a hand in a three-dimensional space. However, there are few restrictions on the operation location, but there is also the trouble of mounting external devices here.

  On the other hand, a user interface system has been developed that does not require a special external device, and can operate the terminal in the air by performing finger recognition using only the resources installed in the portable terminal ( Non-patent document 4). In this system, the troublesomeness of wearing external devices is eliminated, and information is presented to the user on the display of the terminal, so the usage environment is not limited, but because the camera attached to the front of the terminal is used, Factors that make finger recognition difficult. In addition, since the camera lens direction and the user's line-of-sight direction are reversed, when a virtual object linked to the movement of a finger is superimposed, the movement looks unnatural and a realistic operation on three dimensions is performed. There was a problem that it was difficult to realize.

M. Lee, W. Woo, ARKB: 3D vision-based Augmented Reality Keyboard, Int Conf Artif Real Telexistence, vol.13, pp.54--57, 2003. H. Roeber, J. Bacus, C. Tomasi, Typing in Thin Air: The Canesta Projection Keyboard-A New Method of Interaction with Electronic Devices, CHI'03 Extended Abstracts on Human Factors in Computing Systems, pp.712--713, 2003 . P. Mistry, P. Maes, SixthSense: A Wearable Gestural Interface, ACM SIGGRAPH ASIA 2009 Sketches Article, No. 11, 2009. T. Niikura, Y. Hirobe, A. Cassinelli, Y. Watanabe, T. Komuro, M. Ishikawa, In-Air Typing Interface for Mobile Devices with Vibration Feedback, ACM SIGGRAPH 2010 Emerging Technologies Article, No. 15, 2010. C. Harrison, H. Benko, A. D. Wilson, OmniTouch: Wearable MultitouchInteraction Everywhere, Proceedings of the 24th annual ACM symposium on Userinterface software and technology, pp.441--450, 2011

  Therefore, the present invention actually displays a virtual object superimposed on a real image captured by a small camera attached to the back of the mobile terminal, and operates it with fingers placed in the space behind the terminal. The object of the present invention is to propose a user interface system that can be linked to the control of a mobile terminal and the operation of an application in the terminal while feeling as if the user touches the virtual object.

  Therefore, the present invention provides a camera that captures a finger placed in the back space of a mobile terminal from above, a display that displays a virtual keyboard superimposed on the actual image of the finger captured by the camera, and the virtual keyboard. The most important feature is that a virtual key that moves in conjunction with the movement of the finger is provided by operating.

  In the present invention, since the lens direction of the camera and the line-of-sight direction of the operator coincide with each other, the present invention can be easily combined with the AR technology and a more realistic operation can be realized. The inventor used this verification system, and felt the actual key pressed through the keyboard application, and confirmed the effect.

  The biggest feature of this system is that it uses a rear camera instead of a front camera, so the camera lens direction matches the operator's line-of-sight direction, and can be easily combined with AR technology, making it more realistic. Operation can be realized. That is, even if a virtual object is superimposed on a real image photographed by a camera and moved in conjunction with the movement of a finger, unnaturalness is not felt. In addition, by superimposing the finger recognition result on the captured image, when the finger is moved on the virtual object, the user can visually recognize the virtual object as if actually touching it. A certain operation can be realized. Moreover, since a finger does not touch the display directly like a touch panel, sweat or sebum does not adhere to the surface, and the display can be prevented from being soiled.

  When the terminal is operated, it is not necessary to wear a special external device on the body, and information presentation to the user is performed on the display of the terminal. Therefore, there is an advantage that the use environment is not limited. In addition, since the operation can be performed with the elbow bent shallowly, the physical burden on the operator is reduced. Furthermore, since the operator's fingers are placed in the back of the terminal, it is possible to operate with the face close to the terminal display, and even if virtual objects are displayed small on the terminal display, they can be easily seen. Misoperations can be reduced. These are particularly effective for elderly people with poor physical fitness and visual acuity.

  Since a wide space at the back of the terminal serves as the operation area, more convenient terminal operation can be realized. Since the image to be recognized by the finger is captured by the camera on the back of the terminal, the captured image becomes bright because the light from the overhead is not backlit against the camera lens. In addition, the operator's face and indoor fluorescent lamps do not appear in the captured image. Furthermore, texture information of the operator's nails can be used. These advantages facilitate finger recognition. Therefore, the image captured by the camera on the back of the terminal does not include factors that make it difficult to recognize the finger, such as backlight, user's face, and fluorescent light. .

  Since a tracking-based algorithm is used for finger recognition, the stability is improved by increasing the frame rate of the camera.

  As described above, this system is a system that can easily realize a substantial operation by AR in a wide operation space. This system does not require a special operating environment for operation, and the burden on the operator is light and stable finger recognition is realized.

It is a perspective view for demonstrating the mode of operation of the system which implemented this invention. It is a block diagram of the experimental system which implemented this invention. It is the perspective view which attached the camera to the side surface of the terminal lower side as another Example. It is an example of an AR keyboard. It is an example of the fingertip image obtained at the time of fingertip initial detection. It is an example of the virtual key which moves in conjunction with the movement of the finger. It is a display example of the fingertip which made the character of a key visible as another Example.

  Embodiments of the present invention will be described below.

FIG. 1 is a perspective view for explaining a state of operation of a system embodying the present invention.
This system captures the space behind the terminal T where the hand h is placed with the small camera 1 attached to the back of the terminal T, processes the obtained image sequence, and recognizes the movement of the finger f. The virtual keyboard k and the recognition result are superimposed on the image sequence and displayed on the display 2 of the terminal T. Therefore, the operation area is a wide space behind the terminal T, and a more convenient terminal operation can be realized.

FIG. 2 shows a configuration diagram of an experimental system in which the present invention is implemented. FIG. 2A shows a front view thereof, and FIG. 2B shows a rear view thereof.
The experimental system is composed of a small camera 1, a small display 2, and a PC 3.
The small camera 1 is attached to the back surface of the terminal T so as to capture the space behind the display 2. Or you may attach the small camera 1 to the lower side surface of the terminal T, as shown in FIG. As a result, the hand h is not obscured by the terminal T, so that the entire finger can be seen directly and the operability is improved.

  An image sequence including a finger is processed on the PC 3, and the movement of the finger f is recognized as an input to the system. This system can be realized as an actual system by replacing the small display 2 to which the small camera 1 is attached with a portable terminal such as a smartphone and performing the recognition processing performed by the PC 3 by the processing system of the portable terminal.

  Finger movement recognition is performed using a tracking-based algorithm to estimate the position and posture of the fingertip. The recognition process is sequentially performed on the captured image acquired from the camera. The process determines a fingertip parameter estimation initial search point for one acquired photographed image, and estimates the fingertip parameters by template matching. The obtained estimated solution is set as the parameter search initial point of the next photographed image. The fingertip parameters to be estimated are horizontal and vertical translation amounts on a plane perpendicular to the camera optical axis, a change in magnification due to translation in the optical axis direction, and a rotation angle around the optical axis.

  Initial detection of the fingertip is performed by aligning the fingertip with the fingertip contour image arranged at the center of the captured image. The fingertip at the time of initial detection of the fingertip is stored as an image, which is used as a template in subsequent tracking to match the actual image. In the captured image acquired at each time point, a fingertip parameter that minimizes an evaluation function composed of the initial fingertip image and the captured image is obtained. The fingertip image used as a template does not include background information around the finger. Therefore, parameter estimation independent of the background is possible. In addition, the recognition result can be easily superimposed on the actual image.

  In this system, a realistic interface can be presented to the user by superimposing a virtual object that moves in conjunction with the movement of the finger on the captured image. As an example of the interface, an AR keyboard is shown in FIG. In this interface, finger movement and key movement in a three-dimensional space are linked, and a translucent virtual keyboard k is displayed on a real image taken by the camera 1 so that it can be typed. The typed result is displayed in the upper input field N. On the keyboard k, an image obtained by deforming the fingertip f image obtained at the fingertip initial detection based on the parameter of the finger recognition result is displayed. At this time, when the virtual keyboard k is displayed small because the operation area of the terminal T is small, the size of the fingertip f image is displayed in accordance with the size of the virtual keyboard k.

  If the fingertip image is not superimposed, the feeling of touching the virtual keyboard k is not obtained, and the user feels unnatural, but if the fingertip f image is superimposed, the user feels as if he / she is actually touching the virtual keyboard k. Can be visually recognized.

FIG. 5 shows an example of a fingertip image obtained at the initial detection of the fingertip. The key type is determined at a point p near the toe. When this point p is on a certain key, the fingertip size of the finger recognition result at that time is stored, and thereafter, the amount of movement of the key in the depth direction is calculated based on that value. If the fingertip size is smaller than the stored reference value, the amount of movement increases.
Accordingly, if the finger f is moved to the back of the terminal as shown in FIG. 6A to FIG. 6B, the finger size reflected in the camera 1 is reduced, and the key is pushed further to the back. If it is close to 1, the pressed key will also return to its original position. In this way, a feeling that an actual keyboard key is actually pressed with the finger f can be obtained. When the finger f moves and the key type determination point is placed on another key, the key on which the finger was placed just before that is returned to the original position, and the key press reference value is updated.

  The pressing amount for the key on which the finger f is newly placed is calculated according to the new reference value. When a key is pressed beyond a certain threshold, it is determined that the key is input, and information corresponding to the key is fed back to the terminal application and the user. When the key pressing start plane is set to a certain height, when typing a new key, it is necessary to return the finger f to that height, which causes an unstable operation in the air.

  This system avoids this problem by making the key press start plane variable. Since the distance from the camera 1 to the key pressing start plane is not fixed, the user does not have to search for the plane in the air when pressing the key. Since the new key press always starts from the position of the finger f when the finger f is placed on the key, the movement of only the joint at the base of the finger f is performed at that point in time without a large vertical movement of the entire hand h. You can press a new key with a small up and down movement. This enables stable key typing in the air.

  When touching a key with a finger f or the like in key typing, there is a problem that the user cannot accurately recognize the position of the object because the finger f or the like covers the object such as the key. This is often a problem for adults with large fingers or Westerners (so-called Fat Finger Problem). Therefore, as shown in FIG. 7A, the finger f tip is made translucent and displayed so that the character of the key can be seen, or as shown in FIG. 7B, the finger f tip is like an electronic pen. It may be displayed as a thin stick.

  The subjective evaluation in actual operation is described below. In the movement of the finger f on the key, the key pressing start plane is initialized every time the key on which the finger f is placed changes, so that the finger f moves while pushing the key slightly. As a result, the subject could feel as if the finger f was sliding on the actual keyboard. In addition, by making the key press start plane variable, the subject was able to perform key typing with a small vertical movement at an arbitrary position. In addition, the up-and-down movement is the same as the movement of key typing on an actual keyboard. As a result, the subject was able to perform key typing with an actual sense. In this way, the present system can easily present an interface that feels real to the user on the mobile terminal without any unnaturalness.

1 Camera 2 Display 3 PC
f finger h hand k virtual keyboard N input field T terminal

Claims (6)

A camera that shoots the fingers placed in the space behind the mobile device from above,
A display that displays a virtual keyboard superimposed on the actual image of a finger taken with this camera;
By operating this virtual keyboard, virtual keys that move in conjunction with finger movements,
An AR gesture user interface system for mobile terminals, comprising:   2. The AR gesture user interface system for mobile terminals according to claim 1, wherein the camera is attached to a back surface or a lower side surface of the mobile terminal.   The virtual key that moves while interlocking with the movement of the finger decreases the finger size if the finger is moved to the back of the terminal, the virtual key is pushed deeper, and if the finger is brought closer to the camera, the virtual key that is pushed 2. The AR gesture user interface system for portable terminals according to claim 1, wherein the key also returns to the original position.   The pressed amount for the virtual key on which the finger is newly placed is calculated with the finger size stored when the finger is placed as a reference value, and when the key is pushed in exceeding a certain threshold, it is determined as a key input, and the key 2. The AR gesture user interface system for mobile terminals according to claim 1, wherein information corresponding to is fed back to the terminal application and the user.   2. The AR gesture user interface system for portable terminals according to claim 1, wherein when the virtual keyboard is operated, a fingertip is translucently displayed so that characters of the virtual key can be seen.   2. The AR gesture user interface system for portable terminals according to claim 1, wherein when operating the virtual keyboard, a fingertip is deformed and displayed in a thin stick shape like an electronic pen.

Images