JP6197702B2 - Input method, program, and input device - Google Patents

Description

  The present case relates to an input method, a program, and an input device for inputting information to a terminal by movement of the head.

  In recent years, devices (wearable devices) such as computers and gadgets that are supposed to be used while worn on a human body have been developed. These devices employ various UIs (User Interfaces) such as touch input, pen input, and voice input. On the other hand, these UIs do not necessarily provide an operating environment suitable for daily work. For example, in an ICT (Information and Communication Technology) service provided to factories, stores, hospitals, etc., a UI that can operate a device with natural operation without interrupting work related to work is desired.

  Therefore, UI technology that recognizes the posture of the head as an input operation to a computer has been developed. For example, an acceleration sensor is attached to the head to detect the direction of the head, and the computer performs processing corresponding to the direction. In addition, UI technology that detects not only the direction of the head but also the direction of the line of sight and recognizes them as input operations has also been proposed (see Patent Documents 1 and 2).

JP 2012-252568 International Publication No. 2007/105792

  A UI that utilizes head movements can provide a device operating environment with natural movements. However, since the movement of the head may vary depending on the posture at that time and the positional relationship with the visual recognition target, it is difficult to improve the recognition accuracy of the movement. In addition, in understanding the movement state of the head, not only the static state (for example, direction and inclination) but also the dynamic state (for example, rotation amount, movement speed, intensity of movement, etc.) is considered. It is desirable to do. On the other hand, in the existing technology, the former is mainly considered, and the latter is not sufficiently considered, so that it is difficult to improve the recognition accuracy of the operation.

  One of the purposes of the present case was created in view of such a problem, and is to improve the recognition accuracy of the motion. In addition, the present invention is not limited to the above-described purpose, and is an operational effect derived from each configuration shown in “Mode for Carrying Out the Invention” to be described later. Can be positioned as a purpose.

  The disclosed input method is an input method in which information is input to the terminal using the motion of the head. In this input method, the relative inclination of the terminal and the head is acquired, and the operation related to the input of the information is selected based on the inclination. Further, the rotational strength of the head in the selected motion is calculated, and the input section of the motion is determined based on the rotational strength.

  According to the disclosed technique, the recognition accuracy of the operation can be improved.

It is a perspective view which shows the input device which concerns on embodiment. It is a block diagram which illustrates the hardware constitutions of an input device. It is a block block diagram of the program which concerns on input operation It is a graph which illustrates change with time of a pitch angle. It is a figure which shows the pitch angle and yaw angle corresponding to an effective visual field. It is a graph for demonstrating the difference of operation | movement of a terminal and a head, (A) shows the angular velocity fluctuation | variation of a terminal and a head, (B) shows the fluctuation | variation of the difference of angular velocity. It is a graph which shows the fluctuation | variation of rotation intensity [Signal intensity area SMA (AV) about angular velocity]. It is a figure for demonstrating the recognition method of a head gesture, (A) is a figure which shows the movement locus | trajectory of the head gesture on a virtual canvas, (B), (C) is the shape of a gesture, and input operation content. It is a figure which illustrates a relationship. It is an example of the flowchart for specifying the candidate of a gesture area. It is an example of the flowchart for the determination and recognition of a gesture area.

  DESCRIPTION OF EMBODIMENTS Embodiments relating to an input device, an input method, and a program for inputting information to a terminal by head movement will be described below with reference to the drawings. However, the embodiment described below is merely an example, and there is no intention of excluding various modifications and technical applications that are not explicitly described in the embodiment. That is, the present embodiment can be implemented with various modifications (combining the embodiments and each modification) without departing from the spirit of the present embodiment.

[1. Overview of control]
The input device, input method, and program of this embodiment provide a gesture UI to a computer. In this gesture UI, the movement of the head is recognized as an input operation to the computer, and information on the input operation is input to the computer. The computer referred to here includes a portable device with a built-in electronic computer. For example, portable information terminals such as mobile phones, smartphones, Internet terminals, personal computers such as notebook computers and slate computers, music players, cameras, video players, portable game machines, GPS (Global Positioning System) devices, remote control devices, etc. This includes electronic devices (gadgets). Hereinafter, these computers in which input operation information is input are simply referred to as “terminals”.

  In this embodiment, the content of the input operation is interpreted from the relative movement of the head with respect to the terminal, and the information is input to the terminal. Each of the terminal and the head is provided with a measuring device that measures the posture and motion state of the terminal and the head, and the relative movement of the terminal and the head is grasped based on each measurement result. The motion here includes not only a static posture (pose) such as the twist angle and tilt of the head, but also a dynamic posture (motion gesture) that changes over time. In addition, the relative movement here means the movement of the head based on the coordinate system fixed to the terminal (or the movement of the terminal based on the coordinate system fixed to the head). In other words, the motion of the head is grasped as a motion with respect to the terminal, not based on a user's body that is the subject of the motion or a geographical coordinate system fixed to the earth.

  As a measurement device for measuring the movement of the terminal and the head, a triaxial acceleration sensor, a triaxial gyroscope, or an inertial measurement device (IMU, Inertial Measurement Unit) incorporating them is used. The measuring device is provided in each of the terminal and the head. The terminal-side measuring device is fixed and built in the terminal, and the head-side measuring device is fixed and built in the head gadget attached to the head. The head gadget is detachably attached to the head and can be firmly attached to the extent that the head gadget does not fall off due to the movement of the head. The measuring device detects, acquires and calculates parameters corresponding to each operation, and outputs them as time-series data.

In the present embodiment, among the time-series data measured by the measurement device, the data section corresponding to the operation related to the input operation to the terminal is selected in three stages. These selections are performed in real time each time a measurement result is acquired by the measurement device, but can also be performed off-line (non-real time). In this case, the measurement result acquired by the measurement device may be recorded in some data storage device.
The first selection is for selecting an operation related to input of information, and is performed based on at least the relative inclination of the terminal and the head. In the first selection, a candidate for a section in which an operation related to information input is performed is specified. When the first selection condition is satisfied, the information input to the terminal by the gesture of the head is valid (Mode-In state). On the other hand, when the first selection condition is not satisfied, the information input to the terminal by the gesture is not valid (invalidated) (Mode-Out state). In this way, the first selection condition functions like an ON-OFF switch that roughly determines whether or not to refer to the head gesture.

  The second sort is performed subsequent to the first sort and is intended to further sort operations. Here, based on the relative inclination of the head when the first selection condition is satisfied, it is confirmed that the movement range of the line of sight due to the operation is within the effective visual field or the stable focus field. When the second selection condition is satisfied, the data section including the operation at that time is specified as a formal gesture section candidate. The second selection is performed in a state (Mode-In state) in which information input to the terminal is effective by at least the operation of the head. In addition, when the information input to the terminal is not valid (Mode-Out state in which the information input to the terminal is invalidated), the second selection is not performed. The determination of the second selection condition is suspended until at least the first selection condition is satisfied again.

  The third sorting is for determining the input section of motion, and is performed based on the rotational strength of the head when the first and second sorting conditions are satisfied. Here, the rotational strength of the head in the gesture section candidate specified by the first and second selections is determined. As a result, a data section corresponding to a conscious operation (that is, an operation related to input of information) by an operation subject (terminal user) is determined. Thus, the distinction of unconscious movement (involuntary movement) and conscious movement (voluntary movement) to specify the conscious movement is called segmentation. The third sorting condition has a function of accurately discriminating whether or not the head motion is a gesture motion through segmentation.

“Rotational strength” is an index value that represents the intensity of motion in the direction of rotation of the head, and is a constant integral value of the variable corresponding to the rotational motion of the head. Corresponding parameter. This definite integral value is also called a signal intensity area (SMA, Signal Magnitude Area) or a rotation intensity within a certain time.
If the head moves violently in the direction of rotation, the rotational strength increases. Specific examples of rotational intensity include a definite integral value of the head angular velocity, a definite integral value of the head acceleration, a definite integral value of the relative angular velocity of the head relative to the terminal, and a definite integral value of the relative acceleration of the head relative to the terminal. Or a value corresponding to these definite integral values. In the present embodiment, the magnitude of the rotational intensity is determined based on the signal intensity area SMA (AV) for the angular velocity (AV, Angular Velocity) measured by the measuring device. The integration interval is appropriately set within a range of 0.1 to several seconds, for example.

Here, the rotational motion of the head is divided into rotational motions about three axes (x-axis, y-axis, z-axis), and the respective angular velocities are set as ωx (i), ωy (i), and ωz (i). Alternatively, the relative angular velocity of the head relative to the terminal is decomposed into rotational motions about three axes, and the respective angular velocities are set as ωx (i), ωy (i), and ωz (i). The signal intensity area SMA (AV) for each of the three axis angular velocities is the sum of the absolute values of the angular velocities ωx (i), ωy (i), and ωz (i) definitely integrated for a predetermined time T (in the integration interval). Equivalent). I in Equation 1 is an ordinal number given in the order of sampling. The signal intensity area SMA (AV) for the angular velocity is an index representing the intensity of the rotational motion of the head gesture at a predetermined time T when the ordinal number i changes from j to j + T. The head angular velocities are ωx ₁ (i), ωy ₁ (i), and ωz ₁ (i), and the terminal angular velocities are ωx ₂ (i), ωy ₂ (i), and ωz ₂ (i). For example, the signal intensity area SMA (AV) with respect to the relative angular velocity can be expressed as Equation 2.

  The formula for obtaining the signal intensity area SMA (AV) for the angular velocity is not limited to this. For example, considering the effect of each angular velocity ωx (i), ωy (i), ωz (i) on the rotational motion of the head, each definite integral value was multiplied by a predetermined coefficient a, b, c. Things may be calculated (see Equation 3). Alternatively, instead of multiplying the sum of the definite integral values by the predetermined time T, the sum of the definite integral values may be standardized by dividing by the predetermined time T to calculate an index representing the intensity of the rotational motion per unit time. (See Equation 4).

  When the rotational axis of the rotational motion is known, it is possible to obtain the rotational intensity using the acceleration information instead of the angular velocity. For example, since the rotational motion of the head is mainly a rotational motion centered on the neck, the rotational intensity can be calculated using the acceleration of the head instead of the angular velocity of the head. Here, the acceleration of the head is decomposed in three axial directions, and the respective accelerations are set as Ax (i), Ay (i), and Az (i). Alternatively, the relative acceleration of the head with respect to the terminal is decomposed in three axial directions, and the respective accelerations are set as Ax (i), Ay (i), and Az (i). As shown in Equation 5, the signal intensity area SMA (AC) for each of the three axis accelerations is the sum of the absolute values of the respective accelerations Ax (i), Ay (i), and Az (i). It is obtained by multiplying by a predetermined time T (corresponding to the integration interval). It should be noted that the formula for obtaining the signal intensity area SMA (AC) for acceleration is not limited to this, and it is possible to use formulas with various modifications similar to Formula 2, Formula 3, and Formula 4.

  When the data section corresponding to the action related to the input of information is determined, the feature of the action included in the data section is extracted and compared with the preset feature of the gesture, and the gesture corresponding to the action is determined. Be recognized. The feature of the motion is defined based on, for example, the shape and length of the motion trajectory, the velocity vector of each trajectory point constituting the trajectory, the position of each trajectory point, and the like. When the gesture is recognized, input operation information corresponding to the gesture is input to the terminal. A known technique can be applied to a specific gesture recognition method.

[2. Device configuration]
FIG. 1 is a perspective view showing the overall configuration of the input device according to the present embodiment. Here, what controls the information of the input operation to the portable tablet terminal 10 (terminal) using the gesture of the head 30 is illustrated. A head-type head gadget 20 (wearable device) is attached to the head 30 of the user who operates the terminal 10. The head gadget 20 has a function of detecting the movement of the head 30 and transmitting the information to the terminal 10. The gesture of the head 30 is recognized on the terminal 10 side based on information transmitted from the head gadget 20 and is interpreted as an input operation.

The device configuration of the head gadget 20 is illustrated in FIG. Inside the head gadget 20, a CPU 21 (Central Processing Unit), a memory 22, a head IMU 23, and a communication device 24 are provided. These devices operate by receiving power supply from a power source (not shown) (for example, a button battery or a power supply cable).
The CPU 21 is a processor incorporating a control unit (control circuit), an arithmetic unit (arithmetic circuit), a cache memory (register group), and the like. The memory 22 includes a short-term memory element that temporarily stores a program and working data, and a long-term memory element that stores data and programs held for a long time. The former is, for example, ROM (Read Only Memory) and RAM (Random Access Memory), and the latter is non-volatile memory, such as flash memory or EEPROM (Electrically Erasable Programmable Read-Only Memory).

The head IMU 23 is an inertial measurement device that detects the absolute inclination, posture, movement, and the like of the head 30 and measures the direction (orientation), acceleration, and angular velocity of the head 30. The head IMU 23 includes a gyro sensor 23A, an acceleration sensor 23B, and a geomagnetic sensor 23C. The gyro sensor 23 </ b> A detects an angular velocity around three axes (x axis, y axis, z axis) that are coordinate axes fixed to the head 30. The acceleration sensor 23B detects acceleration in three axis directions (x axis, y axis, z axis). The geomagnetic sensor 23C detects the orientation (posture and orientation) of the head 30 with reference to the geographic coordinate system. The data format output from the head IMU 23 is, for example, a rotation matrix format, a quaternion format, an Euler angle format, a raw data (raw data) format, or the like. Hereinafter, data obtained by the head IMU 23 is referred to as “head data”. The head data is transmitted to the CPU 21 and the communication device 24 and recorded on the memory 22 as necessary.
The communication device 24 is a device that performs communication with the terminal 10. The head data obtained by the head IMU 23 is transmitted to the terminal 10 side in real time via the communication device 24. Alternatively, data recorded on the memory 22 is transmitted to the terminal 10 side via the communication device 24. The type of communication is arbitrary, and may be wired communication or wireless communication.

The terminal 10 includes a CPU 11 (processor), a memory 12, a terminal IMU 13, a communication device 14, an RTC 15, a database 16, and a display 17. Preferably, a removable media reading / writing device (not shown) is also provided. These devices operate by receiving power from an arbitrary power source (for example, a built-in battery or a power cable). The CPU 11, memory 12, terminal IMU 13, and communication device 14 are the same devices as the CPU 21, memory 22, head IMU 23, and communication device 24 of the head gadget 20.
The terminal IMU 13 is an inertial measurement device that detects the absolute inclination, posture, movement, and the like of the terminal 10 and measures the direction (orientation), acceleration, and angular velocity of the terminal 10. The terminal IMU 13 includes a gyro sensor 13A, an acceleration sensor 13B, and a geomagnetic sensor 13C. The terminal 10 is based on the angular velocity around the three axes, which are coordinate axes fixed to the terminal 10, the acceleration in the three axis directions, and the geographic coordinate system. Direction (posture, orientation) is detected. Hereinafter, these data are referred to as “terminal data”. The terminal data is transmitted to the CPU 11 and the communication device 14 in real time. The head data obtained by the head IMU 23 is also transmitted to the CPU 11 via the communication device 14. These head data and terminal data are recorded on the memory 12 or a removable medium as necessary.

The RTC 15 (Real Time Clock) is an integrated circuit (chip) for holding and outputting accurate time information regardless of the power state of the terminal 10. Time information output from the RTC 15 is added to all head data and terminal data. Thereby, each of head data and terminal data turns into a time series data group.
The database 16 is a storage device that stores a correspondence relationship between features of an action (gesture) and an input operation corresponding to the gesture. Here, the input operation to the computer is defined in advance for combinations of the shape and length of the locus of motion, the velocity vector of each locus point constituting the locus, the position of each locus point, and the like. The display 17 is a display device that outputs information such as calculation results in the CPU 11, terminal data, and head data. In the present embodiment, the selection stage of the data section related to the motion recognition of the head 30 is displayed. For example, when the first selection condition is satisfied, an image, a symbol, or the like indicating that is displayed on the display 17 to notify the user that the second or third selection condition is being determined. The Note that such notification to the user may be omitted.

[3. program]
FIG. 3 is a block diagram for explaining the processing content (program 1) executed by the CPU 11 of the terminal 10. As shown in FIG. These processing contents are recorded, for example, in the long-term storage element of the memory 12 as an application program, and are expanded and executed in the memory space of the short-term storage element. When the processing contents executed here are functionally classified, the program 1 is provided with an acquisition unit 2, a selection unit 3, a determination unit 4, a locus / feature extraction unit 5, and a real-time collation / classification unit 6.

[3-1. Acquisition Department]
The acquisition unit 2 acquires information regarding the operations of the terminal 10 and the head 30 based on the terminal data transmitted from the terminal IMU 13 and the head data transmitted from the head IMU 23. The acquisition unit 2 includes a head posture estimation unit 2A, a terminal posture estimation unit 2B, and a relative posture estimation unit 2C.
The head posture estimation unit 2A estimates the posture of the head 30 based on the geographic coordinate system based on the head data. The posture (orientation) of the head 30 can be estimated from the geomagnetic direction detected by the geomagnetic sensor 23C, and can also be estimated from acceleration data when the head 30 is stopped. Alternatively, if the influence of offset or noise due to drift is small, the estimation may be performed using the integral value of the angular velocity. The head posture estimation unit 2A calculates the posture of the head 30 in real time using the pitch angle, yaw angle, and roll angle with respect to the geographic coordinate system, adds the time information output from the RTC 15, and transmits the time information to the relative posture estimation unit 2C. To do. Note that the definition of the pitch angle, yaw angle, and roll angle with respect to the geographic coordinate system can be arbitrarily set. In this embodiment, the east-west direction is defined as the rotation axis direction of the pitch angle and the vertical vertical direction is defined as the rotation axis direction of the yaw angle with reference to the rotational motion of the object with the front facing north.

The terminal posture estimation unit 2B estimates the posture of the terminal 10 based on the geographic coordinate system based on the terminal data. The terminal posture estimation unit 2B calculates the posture of the terminal 10 in real time using a pitch angle, a yaw angle, and a roll angle with respect to the geographic coordinate system, adds time information, and transmits the time information to the relative posture estimation unit 2C.
The relative posture estimation unit 2 </ b> C estimates the relative inclination of the terminal 10 and the head 30. Here, relative inclinations (that is, relative pitch angle θ, relative yaw angle ψ, and relative roll angle φ) about the three axes of the pitch direction, the yaw direction, and the roll direction are estimated. The relative posture estimation unit 2C calculates the difference between the tilt of the terminal 10 and the tilt of the head 30 acquired at the same time in real time. Time information is added to the relative inclination information calculated here, and the information is transmitted to the selection unit 3.

[3-2. Sorting unit]
The sorting unit 3 sorts the operation corresponding to the operation input by the gesture based on the relative inclination information acquired by the acquiring unit. The selection unit 3 includes a pitch angle range determination unit 3A, an effective visual field determination unit 3B, and a section candidate identification unit 3C.

The pitch angle range determination unit 3A confirms that the pitch angles of the terminal 10 and the head 30 are appropriate. Here, based on the first selection condition, a data section candidate corresponding to the gesture action is specified. The first selection condition includes at least the following condition A. In the present embodiment, when all of the following conditions A to C are satisfied, it is assumed that the first selection condition is satisfied, and an image, a symbol, or the like indicating that is displayed on the display.
= First selection condition =
A. The relative pitch angle θ is less than a predetermined value B. B. The pitch angle of the terminal 10 is within a predetermined range. The pitch angle of the head 30 is within a predetermined range.

The relationship between these conditions and the pitch angles of the terminal 10 and the head 30 will be described with reference to FIG. The relative pitch angle θ in the condition A corresponds to the difference between the pitch angle of the head 30 indicated by a solid line and the pitch angle of the terminal 10 indicated by a broken line. When the condition A is satisfied, the pitch angle of the head 30 and the pitch angle of the terminal 10 are approximately the same, and the orientation of the head 30 is substantially perpendicular to the display 17 of the terminal 10. Hereinafter, the predetermined value determined here is referred to as a predetermined value P ₀ .

Conditions B and C are for confirming that the pitch angles of the terminal 10 and the head 30 are within a predetermined identification angle range P _{1 to} P ₂ . When the direction of the display 17 is extremely downward or when the direction of the head 30 is extremely upward, these conditions are not satisfied. In FIG. 4, all of the conditions A to C are satisfied in the range of time t _{1 to} t ₂ (input candidate section). In the case where the gesture operation in the posture lying on the sofa or the futon is allowed, the predetermined range (identification angle range P _{1 to} P ₂ ) determined by the condition B and the condition C is expanded, or the condition B , Condition C may be omitted.

  The effective visual field determination unit 3B is for further narrowing down the data section candidates corresponding to the gesture motion in a state where the first selection condition is satisfied by the pitch angle range determination unit 3A. Here, based on the second selection condition, the action corresponding to the operation input by the gesture is selected using the characteristic regarding the gaze action by the human head movement. That is, while the relative pitch angle θ and the relative yaw angle ψ vary within the range corresponding to the effective visual field, the motion is selected as the motion corresponding to the operation input by the gesture. On the other hand, when the fluctuation range of the relative pitch angle θ and the relative yaw angle ψ jumps out of the range corresponding to the effective visual field, it is determined that the operation is no longer the operation corresponding to the operation input by the gesture at that time. .

The effective visual field is an area in which information can be received instantaneously only by eye movement within the visual field. The shape of the effective field of view is, for example, an ellipse that spreads around the gazing point as shown in FIG. 5, has a width of about 30 ° in the major axis direction (horizontal direction), and is about It has a width of 20 °. Following this, the effective visual field determination unit 3B causes the relative pitch angle θ to vary within the first range P _{3 to} P ₄ (for example, within ± 10 °) and the relative yaw angle ψ to be within the second range Y _{1 to} Y. _The movement of the head 30 that fluctuates within ₂ (for example, within ± 15 °) is selected as an action that is effective as a gesture action. The first ranges P _{3 to} P ₄ and the second ranges Y _{1 to} Y ₂ are preferably set as ranges including 0 as shown in FIG.

  FIG. 5 shows a general effective visual field range when looking directly in front. On the other hand, the orientation of the head 30 with respect to the terminal 10 is not necessarily directly in front of the display 17. Therefore, the effective visual field determination unit 3B sets the relative posture in a state where the first selection condition is satisfied by the pitch angle range determination unit 3A as a reference “reference posture”, and a change in the relative posture from the reference posture is determined. It is determined whether it is within the range corresponding to the effective visual field.

The second selection conditions are exemplified below. In the present embodiment, it is assumed that the second selection condition is satisfied when the following conditions D and E are both satisfied.
= Second selection condition =
D. F. Relative pitch angle θ is within the first range with reference posture as a reference. Relative yaw angle ψ fluctuation within the second range with reference posture as the reference (both the first range and the second range correspond to the effective field of view)

  In the second sorting, even if the orientation of the head 30 is not perpendicular to the display 17 of the terminal 10 when the first sorting condition is satisfied, the horizontal direction is ± 15 ° on the basis of the posture, and the vertical direction While the head 30 is moving within a range of ± 10 °, it is determined that the movement is effective as a gesture movement. Such a determination method contributes to both the stricter identification of the gesture section candidate and the natural operation input by the movement of the head 30.

In addition, there exists a stable gaze field that is an area that can be glanced by the movement of the eyeball and the head 30 around the effective visual field. The shape of the stable gazing point is a shape in which an ellipse centered on the gazing point is extended downward, and the region below the gazing point is slightly wider. The minor axis direction (vertical direction) has a width of the third range P _{5 to} P ₆ (for example, about 45 to 70 °), and the major axis direction (horizontal direction) has the fourth range Y _{3 to} Y ₄ (for example, about 60). ~ 90 °) wide. Therefore, when the first selection condition is satisfied, even if the orientation of the head 30 with respect to the terminal 10 is not directly in front of the display 17, the relative pitch angle θ and the relative yaw angle ψ are out of the stable focus field. There is no.

  The section candidate specifying unit 3C specifies a period in which an operation that satisfies both the first and second selection conditions is performed as a “gesture section candidate (motion section candidate)”, and information on angular velocity acquired in this section Is transmitted to the determination unit 4. The gesture section candidate is a period from the time when both the first and second selection conditions are satisfied to the time when either the first or second selection condition is not satisfied. In the first and second sorting, information on the pitch angle and yaw angle is used, whereas in the subsequent third sorting, information on the angular velocity detected by the gyro sensors 13A and 23A is used.

[3-3. Final part]
The determination unit 4 determines an action section recognized as a gesture action from gesture section candidates. The determination unit 4 includes a difference calculation unit 4A, a rotation intensity calculation unit 4B, and a section determination unit 4C.
The difference calculation unit 4A calculates the difference (relative angular velocity) between the angular velocities of the terminal 10 and the head 30 based on the angular velocity information acquired in the section specified as the “gesture section candidate”. Here, for each angular velocity in the pitch direction, yaw direction, and roll direction, the difference between the angular velocity obtained by the terminal IMU 13 and the value obtained by the head IMU 23 is calculated. By calculating such a difference, the offset (initial posture difference) of the gyro sensors 13A and 13B is removed.

  Here, the fluctuation of the angular velocity (yaw angular velocity) in the yaw direction of the terminal 10 and the head 30 is illustrated in FIG. The broken line in the figure corresponds to the pitch angular velocity obtained at the terminal IMU 13, and the solid line corresponds to the pitch angular velocity obtained at the head IMU 23. The difference calculation unit 4A calculates a difference between a broken line and a solid line (a value obtained by subtracting one from the other, or an absolute value thereof) as a relative pitch angular velocity ωx (see FIG. 6B). Similarly, the difference calculation unit 4A calculates a relative roll angular velocity ωy and a relative yaw angular velocity ωz that are differences between the roll angular velocity and the yaw angular velocity. The calculation result here is transmitted to the rotation intensity calculation unit 4B.

  The rotation intensity calculation unit 4B calculates the rotation intensity of the head 30 in the section specified as the “gesture section candidate”. Here, “signal intensity area SMA (AV) for angular velocity” is calculated as the rotation intensity. The predetermined time T related to the calculation of the signal intensity area SMA (AV) is set within a range of 0.1 to 1.0 seconds, for example. As the types of angular velocities, the angular velocities obtained by the head IMU 23 (that is, raw yaw angular velocities, roll angular velocities, pitch angular velocities) may be used, but the relative angular velocities calculated by the difference calculation unit 4A are preferably used. The rotation intensity calculation unit 4B of the present embodiment calculates a signal intensity area SMA (AV) using three types of relative pitch angular velocity ωx, relative roll angular velocity ωy, and relative yaw angular velocity ωz according to the above-described Equations 1 and 2. The information is transmitted to the section determination unit 4C. Since the rotational motion in the roll direction is often weak in the gesture motion of the head 30, the signal strength area SMA (AV) is calculated by omitting the relative roll angular velocity ωy (assuming ωy = 0). Also good.

The section determination unit 4C determines an input section for the gesture motion based on the rotation intensity calculated by the rotation intensity calculation unit 4B. Here, a period in which the value of the signal intensity area SMA (AV) is equal to or greater than the predetermined threshold A ₀ is determined as an input section for gesture motion (that is, a data section corresponding to an operation related to information input). That is, the section determination unit 4C has a function of setting a period during which the rotation intensity is equal to or greater than a predetermined intensity as an input section.
Here, a change in the signal intensity area SMA (AV) of the angular velocity is illustrated in FIG. The input period of the gesture motion is a period from time t _{3 to} t ₄ , time t _{5 to} t ₆ , and time t _{7 to} t ₈ when the signal intensity area SMA (AV) is equal to or greater than the threshold A ₀ . It is determined that the other section is not an input section for the gesture operation even if some operation is performed. By such determination, small noises caused by operations other than gestures are filtered. Thereafter, information on the relative pitch angle θ and the relative yaw angle ψ included in the input section is extracted and transmitted to the trajectory / feature extraction unit 5.

[3-4. Others]
The locus / feature extraction unit 5 calculates a movement locus of the line of sight based on the relative pitch angle θ and the relative yaw angle ψ included in the determined input section. Here, assuming that there is a virtual canvas in front of the head 30, the movement locus of the viewpoint (intersection of the line of sight and the virtual canvas) on the virtual canvas is calculated. In this calculation, information on the relative pitch angle θ and the relative yaw angle ψ included in the input section is recorded in a buffer (not shown) until the gesture in the input section is recognized. Further, the change amount of the relative pitch angle θ is converted into the vertical movement amount of the viewpoint on the virtual canvas, and the change amount of the relative yaw angle ψ is converted into the horizontal movement amount. By these conversions, as shown in FIG. 8A, the movement locus of the line of sight within the input section is drawn. Further, the trajectory / feature extraction unit 5 extracts features of the drawn moving trajectory. The types of features extracted here include the shape and length of the locus, the velocity vector of each locus point constituting the locus, the position of each locus point, and the like. Information on these features is transmitted to the real-time collation / classification unit 6.

The real-time collation / classification unit 6 compares and collates the feature of the movement of the line of sight extracted by the trajectory / feature extraction unit 5 with the feature of the operation (gesture) stored in the database 16, thereby collating the head 30. Recognize gestures. As a specific gesture recognition method, a known pattern recognition method can be employed. For example, the degree of similarity between the features of the movement track of the line of sight is evaluated with respect to all the features of the motion recorded on the database 16, and the motion with the highest similarity is selected.
Further, the real-time collation / classification unit 6 outputs information of an input operation corresponding to the recognized gesture. For example, when the gesture shown in FIG. 8B is recognized, the gesture is interpreted as an operation of “move to the right”. Alternatively, in the case shown in FIG. 8C, it is interpreted as an input operation of the numeral “9”. The information thus interpreted is transmitted to the CPU 11 as an operation input to the terminal 10.

[4. flowchart]
9 and 10 are flowcharts for explaining the procedure of the gesture input method in the terminal 10. These flows correspond to a control procedure by an application program recorded in the memory 12 or a removable medium (not shown), for example, and are read by the CPU 11 and repeatedly executed at a predetermined cycle. The execution period of the program is, for example, a period (for example, 0.01 seconds or less) according to the sampling rate of terminal data and head data in the terminal IMU 13 and head IMU 23.

[4-1. Identifying gesture segment candidates]
The flow in FIG. 9 relates to the identification of gesture section candidates, and the first selection condition and the second selection condition are determined. In this flow, a relative attitude flag F _C and a rotation range flag F _R are used. Relative orientation flag F _C is a flag that is set to F _C = 1 when the first selection condition is satisfied, the rotation range flag F _R F _R when the second selection condition is satisfied This flag is set to = 1. The initial value of these flags is 0.

  In step A1, head data is input from the head IMU 23 to the head posture estimation unit 2A, and terminal data is input from the terminal IMU 13 to the terminal posture estimation unit 2B. Further, the head posture estimation unit 2A performs estimation calculation of the pitch angle, yaw angle, and roll angle of the head 30. Similarly, the terminal posture estimation unit 2B performs estimation calculation of the pitch angle, yaw angle, and roll angle of the terminal 10.

In step A2, the relative posture estimation unit 2C performs estimation calculation of the relative pitch angle θ, the relative yaw angle ψ, and the relative roll angle φ, and acquires the relative inclinations of the terminal 10 and the head 30. In step A3, the pitch angle range determination unit 3A determines whether or not the relative pitch angle θ is less than a predetermined value. This condition because of which becomes even unsatisfied first selection condition in the case of not satisfied, the relative orientation flag F _C proceeds to step A5, the rotation range flag F _R are both set to zero. In the subsequent step A6, the data buffered for gesture recognition described later is erased (cleared), and this flow ends.

Proceeds to step A4 when the condition of step A3 is satisfied, determines whether the pitch angle of each of the terminals 10 and head 30 is within a predetermined range (predetermined identification angular range P ₁ to P in ₂₎ Is done. Even when this condition is not satisfied, the first selection condition is not satisfied, and the process proceeds to step A5. On the contrary, when this condition is satisfied, all of the above conditions A to C are satisfied, and the process proceeds to step A7 because the first selection condition is satisfied.

In step A7, it is determined whether or not the relative posture flag F _C is F _C = 0. This step is a step for determining the timing when the value of the relative posture flag F _C changes from 0 to 1. If F _C = 1 already, steps A8 and A9 are skipped and the process proceeds to step A10. On the other hand, when F _C = 0, the process proceeds to step A8, and the relative attitude flag F _C is set to F _C = 1. In subsequent step A9, the effective visual field determination unit 3B sets the posture at that time to the “reference posture”. For example, if the relative pitch angle θ at that time is the predetermined value θ ₁ and the relative yaw angle ψ is the predetermined value ψ ₁ , these predetermined values θ ₁ and ψ ₁ are stored.

In step A10, the effective visual field determination unit 3B calculates a relative posture based on a difference from the reference posture. Here, how much the relative pitch angle θ varies from the predetermined value θ ₁ and how much the relative yaw angle ψ varies from the predetermined value ψ ₁ is calculated. For example, a value obtained by subtracting the predetermined value θ ₁ from the relative pitch angle θ is calculated as the relative pitch angle difference Δθ, and a value obtained by subtracting the predetermined value ψ ₁ from the relative yaw angle ψ is calculated as the relative yaw angle difference Δψ.

In Step A11, it is determined whether or not the relative pitch angle θ and the relative yaw angle ψ are varied within a range corresponding to the effective visual field with reference to the reference posture. Here, it is determined whether or not the relative pitch angle difference Δθ calculated in the previous step is within the first range (for example, whether P ₃ ≦ Δθ ≦ P ₄ ). Further, it is determined whether or not the relative yaw angle difference Δψ is within the second range (for example, whether Y ₁ ≦ Δψ ≦ Y ₂ ). If these conditions are satisfied, since the second selection condition is also satisfied, the process proceeds to step A12, the rotation range flag F _R is set to F _R = 1. In the subsequent step A13, the current operation is selected as the operation related to the information input operation. That is, the current section is specified as a gesture section candidate, and this flow ends.

On the other hand, if the condition of step A11 is not satisfied, the second selection condition is not satisfied, so the routine proceeds to step A14 where the rotation range flag F _R is set to F _R = 0. In the subsequent step A15, data buffered for gesture recognition described later is erased (cleared), and this flow ends.

[4-2. Confirm gesture section]
The flow of FIG. 10 relates to the determination of the gesture section, and the third selection condition is determined. This flow is executed in parallel with the flow shown in FIG. In step B1, whether relative orientation flag F _C and the rotation range flag F _R is set both to 1 is determined. Here, it is confirmed that the first and second selection conditions are satisfied. When this condition is not satisfied, the process proceeds to step B2, where data buffered for gesture recognition described later is erased (cleared), and this flow ends. On the other hand, when this condition is satisfied, the process proceeds to Step B3.

  In step B3, the difference calculation unit 4A calculates the difference between the angular velocities of the terminal 10 and the head 30 based on the information on the angular velocities acquired in the section specified as the gesture section candidate. For example, a value obtained by subtracting the pitch angular velocity of the head 30 from the pitch angular velocity of the terminal 10 is calculated as the relative pitch angular velocity ωx. Similarly, the relative roll angular velocity ωy and the relative yaw angular velocity ωz are also calculated. In the subsequent step B4, the rotation intensity calculation unit 4B calculates the rotation intensity SMA (VA) for the angular velocity. Here, for example, the signal intensity area SMA (AV) of the angular velocity is calculated as the rotation intensity according to Expressions 1 and 2.

In step B5, whether the angular velocity signal intensity area SMA (AV) is the threshold value A ₀ or not is determined. Here, when SMA (AV) ≧ A ₀ , the process proceeds to step B6, and the section determining unit 4C determines that the movement of the head 30 at that time is included in the gesture section. That is, at least the motion at that time is determined to be a part of the gesture motion. In the subsequent step B7, information on the relative pitch angle θ and the relative yaw angle ψ at that time is transmitted to the trajectory / feature extraction unit 5 and recorded in the buffer. In step B8, the buffer flag F _B indicating that the data related to the gesture is buffered is set to F _B = 1, and this flow ends. As long as the first and second selection conditions are satisfied and the angular velocity signal intensity area SMA (AV) is equal to or greater than the threshold A ₀ , information on the relative pitch angle θ and the relative yaw angle ψ is obtained as the locus / feature extraction unit 5. Will be buffered one by one.

When SMA (AV) <A ₀ in step B5, the process proceeds to step B9, and it is determined whether or not the buffer flag F _B is F _B = 1. If this condition is not satisfied, the information about the relative pitch angle θ and the relative yaw angle ψ has not been buffered yet, so this flow is terminated as it is. This corresponds to, for example, a state in which the angular velocity signal intensity area SMA (AV) is less than the threshold A ₀ immediately after the first and second selection conditions are satisfied.

On the other hand, when the condition of step B9 is satisfied, the process proceeds to step B10. In step B10, the trajectory / feature extraction unit 5 calculates the movement trajectory of the line of sight based on the buffered information on the relative pitch angle θ and the relative yaw angle ψ, and extracts the trajectory features. In addition, the real-time collation / classification unit 6 compares and collates the characteristics of the movement trajectory with the characteristics of the operation (gesture) on the database 16, recognizes the gesture of the head 30, and performs an input operation corresponding to the gesture. Is output.
In step B11 where the recognition and output of the gesture are completed, the relative posture flag F _C , the rotation range flag F _R , and the buffer flag F _B are all reset to 0. In step B12, the buffer of the trajectory / feature extraction unit 5 is erased (cleared) in preparation for a new gesture input thereafter.

[5. Action, effect]
(1) In the above embodiment [the terminal 10 (input device), the gesture input method executed in the terminal 10 (the flowchart shown in FIGS. 9 and 10) and the program 1 relating to the gesture input], information input When the “data section (that is, the input section of the gesture motion)” in which such a motion is performed is determined, the “gesture section candidate” is specified based on the relative inclination of the terminal 10 and the head 30. In addition, the final “data section” is determined based on the rotation intensity of the head 30 in the identified “gesture section candidate”.
In other words, the relative inclination information is used to identify the “gesture section candidate”, but when further narrowing down and confirming the candidate, the rotation intensity indicating the intensity of the rotational motion of the head 30 is used. Information is used.

By adopting such a method, it is possible to accurately grasp a section in which an arbitrary gesture operation has been performed, and to improve the segmentation accuracy. As a result, the features of the movement trajectory extracted by the trajectory / feature extraction unit 5 are optimized, and the collation accuracy in the real-time collation / classification unit 6 is improved. Therefore, the recognition accuracy of gesture motion can be improved.
For example, conscious voluntary movement by the user and involuntary movement that occurs unconsciously can be distinguished with high accuracy, and the probability of erroneous input can be greatly reduced. Further, even a natural small movement of the head 30 that does not give the user a sense of restraint can be determined as a gesture related to an input operation.

  (2) In the above-described embodiment, as shown in the condition A, an operation in which the relative pitch angle θ between the terminal 10 and the head 30 is less than a predetermined value is selected. By specifying the data section candidate using the relative pitch angle θ, the operation of the head 30 relative to the terminal 10 can be accurately grasped. For example, a gesture operation in a posture lying on a sofa or a futon can be allowed, and the motion identification accuracy can be improved.

  (3) In the above embodiment, as shown in the condition B and the condition C, the operations in which the pitch angles of the terminal 10 and the head 30 are within a predetermined range are selected. By restricting the angle range of each pitch angle, it is possible to exclude an inclination unrelated to the operation, and to improve the identification accuracy of the operation. For example, as shown in FIG. 4, an operation when the display 17 is extremely downward or when the head 30 is extremely upward can be excluded from recognition targets. Further, in these states, the first selection condition is not satisfied, and data related to gesture recognition is not buffered, so that the calculation load can be reduced.

  (4) In the above embodiment, as shown in the conditions D and E, the operations are selected on the condition that the fluctuations in the relative pitch angle θ and the relative yaw angle ψ are within the range corresponding to the effective visual field. The effective visual field is an area in which information can be received instantaneously only by eye movement within the visual field. That is, it can be said that a gesture accompanied by a head movement exceeding the effective visual field is an unnatural movement that cannot receive information instantaneously only by eye movement. In the above embodiment, such an unnatural motion is excluded from the recognition target, so that it is possible to realize an operation input by natural movement of the head 30 while more precisely identifying a gesture section candidate.

  (5) Under the above conditions D and E, changes in the relative pitch angle θ and the relative yaw angle ψ with respect to the reference posture are determined. The reference posture is a relative posture in a state where the first selection condition is satisfied. That is, in the determination of condition D and condition E, the center position of the effective visual field is changed according to the relative positional relationship between the terminal 10 and the head 30 immediately before the determination is made. Thereby, even if the orientation of the head 30 is not perpendicular to the display 17 of the terminal 10, it is possible to set an effective visual field based on the tilted viewing posture. Therefore, it is possible to realize an operation input by natural movement of the head 30 while more precisely identifying gesture section candidates.

(6) Under the above conditions D and E, the variation of the relative pitch angle θ is within the first range P _{3 to} P ₄ and the variation of the relative yaw angle ψ is within the second range Y _{1 to} Y ₂ . It is determined whether or not there is. As described above, by performing the second selection on the condition that the relative pitch angle θ and the relative yaw angle ψ are relatively small, it is possible to more precisely identify the gesture section candidate and to identify the motion identification accuracy. Can be improved.
(7) In the above-described embodiment, in the section determining unit 4C, a period in which the rotation intensity is equal to or greater than the predetermined intensity is determined as the gesture motion input section. In this way, by using the intensity of motion in the rotation direction, it is possible to accurately grasp the conscious motion of the head 30 and improve the motion identification accuracy.

(8) In the above embodiment, as shown in Equations 1 and 2, the signal intensity area SMA (AV) for the angular velocity corresponding to the rotational strength of the head is calculated. By adopting such a calculation method, it is possible to quantitatively evaluate the intensity of the rotational motion of the head 30 about all three axes, and to improve the motion identification accuracy.
(9) In the above embodiment, the signal intensity area SMA (AV) is calculated using the relative angular velocity calculated by the difference calculation unit 4A. As a result, it is possible to remove the influence of body motion noise (body motion artifact, noise caused by walking motion, vehicle shake, etc.) and improve motion identification accuracy.

[6. Modified example]
Regardless of an example of the disclosed embodiment, various modifications can be made without departing from the spirit of the present embodiment. Each structure and each process of this embodiment can be selected as needed, or may be combined suitably.

  In the above-described embodiment, as shown in FIG. 1, a headset-type wearable device is illustrated, but the method for acquiring the movement of the head 30 is not limited to this. For example, an ear-hook type or glasses-type wearable device in which the head IMU 23 is incorporated may be used. In addition, it is good also as a control structure which calculates the relative inclination of the terminal 10 and the head 30, and the rotational intensity of the head 30 by analyzing the image image | photographed with the built-in camera of the terminal 10. FIG.

  Further, in the above-described embodiment, an example of recognizing a gesture inside the terminal 10 has been illustrated, but an entity that performs control related to gesture recognition is arbitrary. For example, the above program 1 may be executed inside the head gadget 20 or may be executed on a computer other than the terminal 10 and the head gadget 20. In this case, it is preferable that the subject performing the control can communicate with the terminal 10 and the head gadget 20 by wired or wireless communication.

  In the above-described embodiment, the program 1 shown in FIG. 3 is recorded on the memory 12 of the terminal 10, but the recording medium on which the program 1 is recorded is not limited to this. For example, it may be provided in a form recorded on a computer-readable recording medium such as a magnetic recording medium or an optical recording medium. In this case, the computer reads the program from the recording medium, transfers it to the internal storage device or the external storage device, stores it, and executes the program. In the above-described embodiment, the functions shown in FIG. 3 are implemented on software. However, some or all of these functions may be provided as hardware (logic circuit).

[7. Addendum]
The following additional remarks are disclosed with respect to the embodiment including the above modification.
[7-1. input method]
(Appendix 1)
In the input method to input information to the terminal using the movement of the head,
Obtaining a relative inclination of the terminal and the head;
Based on the inclination, the operation related to the input of the information is selected,
Calculate the rotational strength of the head in the selected movement,
An input method for determining an input section of the operation based on the rotation intensity.
(Appendix 2)
The input method according to appendix 1, wherein a relative rotational intensity of the head with respect to the terminal is calculated.

(Appendix 3)
The input method according to appendix 1 or 2, characterized in that the movement is performed when the relative pitch angle between the terminal and the head is less than a predetermined value.
(Appendix 4)
The input method according to any one of appendices 1 to 3, wherein the movements in which the pitch angles of the terminal and the head are within a predetermined range are selected.

(Appendix 5)
5. The operation according to any one of appendices 1 to 4, wherein the operation is performed in which the relative pitch angle and yaw angle variation of the terminal and the head are within a range corresponding to an effective visual field. input method.
(Appendix 6)
The input method according to claim 5, wherein a reference position of the effective visual field is set based on a relative pitch angle and yaw angle between the terminal and the head.

(Appendix 7)
The movement is selected in which the relative pitch angle variation between the terminal and the head is within a first range, and the relative yaw angle variation between the terminal and the head is within a second range. The input method according to appendix 5 or 6, characterized by the above.
(Appendix 8)
The input method according to any one of appendices 1 to 7, wherein a period in which the rotational intensity is equal to or greater than a predetermined intensity is set as the input section.

〔ω_x(i), ω_y(i), ω_z(i) は三軸それぞれの角速度，Tは所定時間〕 (Appendix 9)
The signal intensity area for the angular velocity of the head (that is, a parameter corresponding to a definite integral value of the angular velocity of the head) is calculated as the rotational intensity. The input method described.
(Appendix 10)
The input method according to appendix 9, wherein the signal intensity area is calculated using the following formula 1.

[Ω _x (i), ω _y (i), ω _z (i) are the angular velocities of each of the three axes, and T is the predetermined time]

[7-2. Input device]
(Appendix 11)
An input device for inputting information to the terminal using the movement of the head,
A processor;
Memory,
An acquisition unit for acquiring a relative inclination of the terminal and the head;
Based on the inclination acquired by the acquisition unit, a selection unit that selects an operation related to the input of the information;
A calculation unit for calculating the rotational strength of the head in the movement selected by the selection unit;
A determination unit for determining an input section of the motion based on the rotation intensity calculated by the calculation unit;
An input device comprising:
(Appendix 12)
The input device according to appendix 11, wherein the calculation unit calculates a relative rotational intensity of the head with respect to the terminal.

(Appendix 13)
13. The input device according to appendix 11 or 12, wherein the selecting unit selects the operation in which a relative pitch angle between the terminal and the head is less than a predetermined value.
(Appendix 14)
The input device according to any one of appendices 11 to 13, wherein the sorting unit sorts the movement in which each pitch angle of the terminal and the head is within a predetermined range.

(Appendix 15)
Any one of appendixes 11 to 14, wherein the selecting unit selects the operation in which the relative pitch angle and yaw angle of the terminal and the head are within a range corresponding to an effective visual field. The input device according to item 1.
(Appendix 16)
The input device according to appendix 15, wherein the selecting unit sets a reference position of the effective visual field based on a relative pitch angle and yaw angle between the terminal and the head.

(Appendix 17)
The selection unit has a variation in relative pitch angle between the terminal and the head within a first range, and a variation in relative yaw angle between the terminal and the head within a second range. The input device according to appendix 15 or 16, wherein the operation is selected.
(Appendix 18)
The input device according to any one of appendices 11 to 17, wherein the determination unit sets a period in which the rotational intensity is equal to or greater than a predetermined intensity as the input section.

(Appendix 19)
The calculation unit calculates a signal intensity area for the angular velocity of the head (that is, a parameter corresponding to a constant integral value of the angular velocity of the head) as the rotation intensity, The input device according to any one of the above.
(Appendix 20)
20. The input device according to appendix 19, wherein the calculation unit calculates the signal intensity area using Equation 1 above.

[7-3. Program (recording medium)]
(Appendix 21)
In a program (or a computer-readable recording medium in which the program is recorded) for inputting information to the terminal using the movement of the head,
Obtaining a relative inclination of the terminal and the head;
Based on the inclination, the operation related to the input of the information is selected,
Calculate the rotational strength of the head in the selected movement,
A program (or a computer-readable recording medium on which a program is recorded) that causes a computer to execute processing for determining an input interval of the operation based on the rotation intensity.
(Appendix 22)
The program according to appendix 21, which causes a computer to execute a process of calculating the relative rotational intensity of the head with respect to the terminal (or a computer-readable recording medium on which the program is recorded).

(Appendix 23)
23. The program according to appendix 21 or 22, which causes a computer to execute a process of selecting the operation in which the relative pitch angle between the terminal and the head is less than a predetermined value (or a computer-readable recording medium on which the program is recorded).
(Appendix 24)
24. The program according to any one of appendices 21 to 23, which causes a computer to execute a process of selecting the operation in which the pitch angles of the terminal and the head are within a predetermined range [or computer-readable recording the program Recording medium].

(Appendix 25)
25. The supplementary note 21 to 24, which causes the computer to execute a process of selecting the movement in which the relative pitch angle and yaw angle of the terminal and the head are within a range corresponding to an effective visual field. Program [or computer-readable recording medium recording the program].
(Appendix 26)
The program according to appendix 25, which causes a computer to execute a process of setting a reference position of the effective visual field based on a relative pitch angle and yaw angle of the terminal and the head [or a computer-readable record recording the program Medium].

(Appendix 27)
The movement is selected in which the relative pitch angle variation between the terminal and the head is within a first range, and the relative yaw angle variation between the terminal and the head is within a second range. 27. The program according to appendix 25 or 26, which causes a computer to execute processing (or a computer-readable recording medium on which the program is recorded).
(Appendix 28)
28. The program according to any one of appendices 21 to 27, which causes a computer to execute a process in which a period in which the rotational intensity is equal to or greater than a predetermined intensity is the input section [or a computer-readable recording medium on which the program is recorded].

(Appendix 29)
Any one of appendices 21 to 28 that causes a computer to execute a process of calculating a signal intensity area for the angular velocity of the head (that is, a parameter corresponding to a definite integral value of the angular velocity of the head) as the rotational intensity. The described program [or a computer-readable recording medium on which the program is recorded].
(Appendix 30)
30. The program according to appendix 29 [or a computer-readable recording medium on which the program is recorded] that causes a computer to execute the process of calculating the signal intensity area using Equation 1 above.

1 program (gesture recognition program)
2 acquisition unit 2A head posture estimation unit 2B terminal posture estimation unit 2C relative posture estimation unit 3 selection unit 3A pitch angle range determination unit 3B effective visual field determination unit 3C section candidate specification unit 4 determination unit 4A difference calculation unit 4B rotation intensity calculation unit (Calculation unit)
4C section confirmation part (confirmation part)
5 Trajectory / feature extraction unit 6 Real-time collation / classification unit 10 Terminal 13 Terminal IMU
20 Head Gadget 23 Head IMU
30 heads

Claims (10)

In the input method to input information to the terminal using the movement of the head,
Obtaining a relative inclination of the terminal and the head;
Based on the inclination, the operation related to the input of the information is selected,
Calculate the rotational strength of the head in the selected movement,
An input method for determining an input section of the operation based on the rotation intensity. The input method according to claim 1, wherein a relative rotation intensity of the head with respect to the terminal is calculated. The input method according to claim 1, wherein the movement is performed when a relative pitch angle between the terminal and the head is less than a predetermined value. The input method according to claim 1, wherein the movements in which the pitch angles of the terminal and the head are within a predetermined range are selected. The said operation | movement which the fluctuation | variation of the relative pitch angle of the said terminal and the said head and a yaw angle is in the range corresponding to an effective visual field is selected, The any one of Claims 1-4 characterized by the above-mentioned. Input method. The input method according to claim 5, wherein a reference position of the effective visual field is set based on a relative pitch angle and yaw angle between the terminal and the head. The movement is selected in which the relative pitch angle variation between the terminal and the head is within a first range, and the relative yaw angle variation between the terminal and the head is within a second range. The input method according to claim 5 or 6, characterized in that: The input method according to claim 1, wherein a period in which the rotational intensity is equal to or greater than a predetermined intensity is set as the input section. In a program that uses head movements to input information into the terminal,
Obtaining a relative inclination of the terminal and the head;
Based on the inclination, the operation related to the input of the information is selected,
Calculate the rotational strength of the head in the selected movement,
A program that causes a computer to execute a process of determining an input interval of the operation based on the rotation intensity. In an input device that inputs information to the terminal using the motion of the head,
An acquisition unit for acquiring a relative inclination of the terminal and the head;
Based on the inclination acquired by the acquisition unit, a selection unit that selects an operation related to the input of the information;
A calculation unit for calculating the rotational strength of the head in the movement selected by the selection unit;
A determination unit for determining an input section of the motion based on the rotation intensity calculated by the calculation unit;
An input device comprising:

Images