Virtual keyboard for a non-tactile three dimensional user interface

Abstract

A method, including presenting, by a computer system executing a non-tactile three dimensional user interface, a virtual keyboard on a display, the virtual keyboard including multiple virtual keys, and capturing a sequence of depth maps over time of a body part of a human subject. On the display, a cursor is presented at positions indicated by the body part in the captured sequence of depth maps, and one of the multiple virtual keys is selected in response to an interruption of a motion of the presented cursor in proximity to the one of the multiple virtual keys.

Description

FIELD OF THE INVENTION

This invention relates generally to user interfaces for computerized systems, and specifically to user interfaces that are based on three-dimensional sensing.

BACKGROUND OF THE INVENTION

Many different types of user interface devices and methods are currently available. Common tactile interface devices include a computer keyboard, a mouse and a joystick. Touch screens detect the presence and location of a touch by a finger or other object within the display area. Infrared remote controls are widely used, and “wearable” hardware devices have been developed, as well, for purposes of remote control.

Computer interfaces based on three-dimensional (3D) sensing of parts of a user's body have also been proposed. For example, PCT International Publication WO 03/071410, whose disclosure is incorporated herein by reference, describes a gesture recognition system using depth-perceptive sensors. A 3D sensor, typically positioned in a room in proximity to the user, provides position information, which is used to identify gestures created by a body part of interest. The gestures are recognized based on the shape of the body part and its position and orientation over an interval. The gesture is classified for determining an input into a related electronic device.

Documents incorporated by reference in the present patent application are to be considered an integral part of the application except that to the extent any terms are defined in these incorporated documents in a manner that conflicts with the definitions made explicitly or implicitly in the present specification, only the definitions in the present specification should be considered.

As another example, U.S. Pat. No. 7,348,963, whose disclosure is incorporated herein by reference, describes an interactive video display system, in which a display screen displays a visual image, and a camera captures 3D information regarding an object in an interactive area located in front of the display screen. A computer system directs the display screen to change the visual image in response to changes in the object.

SUMMARY OF THE INVENTION

There is provided, in accordance with an embodiment of the present invention a method, including presenting, by a computer system executing a non-tactile three dimensional user interface, a virtual keyboard on a display, the virtual keyboard including multiple virtual keys, capturing a sequence of depth maps over time of a body part of a human subject, presenting, on the display, a cursor at positions indicated by the body part in the captured sequence of depth maps, and selecting one of the multiple virtual keys in response to an interruption of a motion of the presented cursor in proximity to the one of the multiple virtual keys.

There is also provided, in accordance with an embodiment of the present invention an apparatus, including a display, and a computer executing a non-tactile three dimensional user interface and configured to present a virtual keyboard on a display, the virtual keyboard including multiple virtual keys, to capture a sequence of depth maps over time of a body part of a human subject, to present, on the display, a cursor at positions indicated by the body part in the captured sequence of depth maps, and to select one of the multiple virtual keys in response to an interruption of a motion of the presented cursor in proximity to the one of the multiple virtual keys.

There is further provided, in accordance with an embodiment of the present invention a computer software product including a non-transitory computer-readable medium, in which program instructions are stored, which instructions, when read by a computer executing a non-tactile three dimensional user interface, cause the computer present a virtual keyboard on a display, the virtual keyboard including multiple virtual keys, to capture a sequence of depth maps over time of a body part of a human subject, to present, on the display, a cursor at positions indicated by the body part in the captured sequence of depth maps, and to select one of the multiple virtual keys in response to an interruption of a motion of the presented cursor in proximity to the one of the multiple virtual keys.

BRIEF DESCRIPTION OF THE DRAWINGS

The disclosure is herein described, by way of example only, with reference to the accompanying drawings, wherein:

FIG. 1 is a schematic pictorial illustration of a system configured to implement a virtual keyboard for a non-tactile three dimensional user interface, in accordance with an embodiment of the present invention;

FIG. 2 is a flow diagram that schematically illustrates a method of interacting with the virtual keyboard, in accordance with an embodiment of the present invention;

FIG. 3 is a schematic pictorial illustration of the virtual keyboard, in accordance with an embodiment of the present invention;

FIG. 4 is a schematic pictorial illustration of character input via the virtual keyboard, in accordance with an embodiment of the present invention; and

FIG. 5 is a schematic pictorial illustration of character input via the virtual keyboard, in accordance with an alternative embodiment of the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Overview

Computer keyboards typically comprise an arrangement of physical keys which act electronic switches. Despite the development of alternative input devices such as mice, touchscreens and pen devices, computer keyboards remain a commonly used versatile device for direct input into computers.

When using a tactile input device such as a computer keyboard, a user typically presses the physical keys in order to convey alphanumeric text and system commands (e.g., an Enter key or cursor keys) to a computer coupled to the keyboard. However, when interacting with a non-tactile 3D user interface (also referred to herein as a 3D user interface), the user may perform gestures in mid-air, and perform the gestures from different positions within a field of view of a 3D sensor coupled to the 3D user interface.

Embodiments of the present invention provide methods and systems for conveying input to a non-tactile 3D user interface via a virtual keyboard presented on a display. The virtual keyboard may comprise multiple virtual keys that represent alphanumeric characters (i.e., “A”-“Z” and “0”-“9”), symbol characters (e.g., “@” and “+”), punctuation characters and control commands (e.g., an Enter key, and cursor and function keys). The virtual keyboard may also comprise a box that is configured to present any text or other characters that were input by the user via the virtual keyboard. In the description and in the claims, the term “virtual keyboard” is to be understood as a graphic representation of a keyboard that does not operate tactilely, and is presented on a display.

The 3D user interface can be configured to track the user's hand (or any other limb), and to position a cursor on a display at positions indicated by the hand's position. In one embodiment, the user can input a given virtual key by keeping the hand relatively steady as the cursor is presented over the given virtual key for a specified time period. In an additional embodiment, the specified time period may be shortened if a language model indicates that given virtual key is predicted based on previously entered virtual keys. For example, if the user previously entered the letters “bl”, and then positions the cursor over the virtual key “i”, the 3D user interface may accept the letter “i” after the cursor is presented in proximity to the virtual key “i” for a shorter specified time period, (e.g., 0.2 seconds). The 3D user interface can accept the letter “i” after the shorter time period since the language model can identify “bli” as first characters in the words “blink”, “blind”, etc. However, if after entering the letters “bl”, the user positions the cursor over the virtual key “z”, then the 3D user interface may accept “z” after the cursor is positioned over the virtual key “z” for a longer specified time period (e.g., one second).

In an alternative embodiment, as the user makes a smooth change of direction of a trajectory of the hand, the 3D user interface can apply a language model to select a given virtual key that the user intended to input. For example, if the letters “bac” were previously input by the user, and the user changes the direction of the hand's trajectory as the cursor is presented in the vicinity of virtual keys “i”, “o”, “j” and “k”, the language model can select the letter “k”, thereby completing the word “back”.

Utilizing a language model can provide a best guess of the user's intended input that enables the user to enter characters (i.e., via the virtual keyboard) more rapidly. Additionally, the smooth change of direction is natural during fast text input, and may have ergonomic advantages.

System Description

FIG. 1 is a schematic, pictorial illustration of a non-tactile 3D user interface 20 (also referred to herein as the 3D user interface) for operation by a user 22 of a computer 26, in accordance with an embodiment of the present invention. The non-tactile 3D user interface is based on a 3D sensing device 24 coupled to the computer, which captures 3D scene information of a scene that includes the body or at least a body part, such as a hand 30, of the user. Device 24 or a separate camera (not shown in the figures) may also capture video images of the scene. The information captured by device 24 is processed by computer 26, which drives a display 28 accordingly.

Computer 26, executing 3D user interface 20, processes data generated by device 24 in order to reconstruct a 3D map of user 22. The term “3D map” refers to a set of 3D coordinates measured with reference to a generally horizontal X-axis 32, a generally vertical Y-axis 34 and a depth Z-axis 36, based on device 24. The set of 3D coordinates can represent the surface of a given object, in this case the user's body. In operation, user 22 moves hand 30 in an X-Y plane 38 to interact with a virtual keyboard 40 and a cursor 42, which are both presented on the display.

In one embodiment, device 24 projects a pattern of spots onto the object and captures an image of the projected pattern. Computer 26 then computes the 3D coordinates of points on the surface of the user's body by triangulation, based on transverse shifts of the spots in the pattern. Methods and devices for this sort of triangulation-based 3D mapping using a projected pattern are described, for example, in PCT International Publications WO 2007/043036, WO 2007/105205 and WO 2008/120217, whose disclosures are incorporated herein by reference. Alternatively, interface 20 may use other methods of 3D mapping, using single or multiple cameras or other types of sensors, as are known in the art.

Computer 26 is configured to capture, via 3D sensing device 24, a sequence of depth maps over time. Each of the depth maps comprises a representation of a scene as a two-dimensional matrix of pixels, where each pixel corresponds to a respective location in the scene, and has a respective pixel depth value that is indicative of the distance from a certain reference location to the respective scene location. In other words, pixel values in the depth map indicate topographical information, rather than a brightness level and/or a color of any objects in the scene. For example, depth maps can be created by detecting and processing an image of an object onto which a laser speckle pattern is projected, as described in PCT International Publication WO 2007/043036 A1, whose disclosure is incorporated herein by reference.

In some embodiments, computer 26 can process the depth maps in order to segment and identify objects in the scene. Specifically, computer 26 can identify objects such as humanoid forms (i.e., 3D shapes whose structure resembles that of a human being) in a given depth map, and use changes in the identified objects (i.e., from scene to scene) as input for controlling computer applications.

For example, PCT International Publication WO 2007/132451, whose disclosure is incorporated herein by reference, describes a computer-implemented method where a given depth map is segmented in order to find a contour of a humanoid body. The contour can then be processed in order to identify a torso and one or more limbs of the body. An input can then be generated to control an application program running on a computer by analyzing a disposition of at least one of the identified limbs in the captured depth map.

In some embodiments, computer 26 can process captured depth maps in order to track a position of hand 30. By tracking the hand position, 3D user interface 20 can use hand 30 as a pointing device in order to control the computer or other devices such as a television and a set-top box. Additionally or alternatively, 3D user interface 20 may implement “digits input”, where user 22 uses hand 30 as a pointing device to select a digit presented on display 28. Tracking hand points and digits input are described in further detail in PCT International Publication WO IB2010/051055.

Computer 26 typically comprises a general-purpose computer processor, which is programmed in software to carry out the functions described hereinbelow. The software may be downloaded to the processor in electronic form, over a network, for example, or it may alternatively be provided on non-transitory tangible media, such as optical, magnetic, or electronic memory media. Alternatively or additionally, some or all of the functions of the image processor may be implemented in dedicated hardware, such as a custom or semi-custom integrated circuit or a programmable digital signal processor (DSP). Although computer 26 is shown in FIG. 1, by way of example, as a separate unit from sensing device 24, some or all of the processing functions of the computer may be performed by suitable dedicated circuitry within the housing of the sensing device or otherwise associated with the sensing device.

As another alternative, these processing functions may be carried out by a suitable processor that is integrated with display 28 (in a television set, for example) or with any other suitable sort of computerized device, such as a game console or media player. The sensing functions of device 24 may likewise be integrated into the computer or other computerized apparatus that is to be controlled by the sensor output.

Virtual Keyboard Interaction

FIG. 2 is a flow diagram that schematically illustrates a method of character input using virtual keyboard 40, in accordance with an embodiment of the present invention, and FIG. 3 is a schematic pictorial illustration of the virtual keyboard, in accordance with an embodiment of the present invention. FIG. 4 is a schematic pictorial illustration of user 22 interacting with virtual keyboard 40, in accordance with the embodiment of the present invention, and FIG. 5 is a schematic pictorial illustration of the user interacting with the virtual keyboard, in accordance with an alternative embodiment of the present invention.

In a presentation step 50 in the flow diagram, 3D user interface 20 presents virtual keyboard 40 on display 28. In the configuration shown in FIG. 3, virtual keyboard 40 comprises virtual keys 70, which can be presented with a surrounding border 72. As user 22 selects one of virtual keys 70, which is alphanumeric, using the embodiments described herein, 3D user interface 20 can present the corresponding alphanumeric character in a text box 74.

Virtual keys 70 may comprise alphanumeric characters, a backspace key, a space bar, symbols and punctuation (e.g., “@” and “?”). Additionally virtual keys 70 may include control keys (e.g., an Enter key and cursor keys) and function keys (e.g., F1, F2, F3, etc.). In some embodiments, 3D user interface 20 can toggle the virtual keys between different modes (e.g., upper and lower case characters) and character sets (e.g., English, Arabic, Chinese and Hebrew). Additionally, the design of virtual keyboard 40 may include “empty” areas 76 between each of the virtual keys, so that user 22 can easily direct cursor 42 to an empty location, thereby reducing the probability of a false positive input.

Returning to the flow diagram, in an initialization step 52, computer 26 sets initial values for a standard time period and an override time period that can be used to by the 3D user interface for deciding when to accept a keystroke on virtual keyboard 40, as described in further detail hereinbelow. Typically, the standard time period is shorter than the override time period, and are both stored as parameters in 3D user interface 20.

In some embodiments, 3D user interface 20 can automatically adjust the standard and the override time periods in response to a proficiency of user 22. In other words, 3D user interface 20 can initially set the standard and the override time periods period to first values, and then modify the standard and the override time periods according to a skill level of user 22. In an embodiment, 3D user interface 20 may measure the user's skill level by calculating an average time interval that is required for the user to transition from a first given virtual key 70 to a second given virtual key 70 (e.g., from “a” to “t”). For example, for every five alphanumeric inputs (i.e., via the virtual keyboard) computer 26 can calculate the average time period between each of the inputs and classify the user's skill level to one of several (e.g., three) levels, where each of the levels is associated with different standard and override time period parameters.

Additionally or alternatively, 3D user interface 20 can adjust the specified time period using factors such as:

• • A number of typing errors user 22 made using virtual keyboard 40. The number of errors may be recorded according to the number of times a “Backspace” virtual key is selected. For example, if user 22 has a high error rate, 3D user interface 20 can increase the specified time period.
  • 3D user interface 20 can present a given virtual key 70 that is dedicated to changing the specified time period (e.g., a “+” or a “−” key).
  • A unique specified time period that is associated with each user of 3D user interface 20.
  • A profile of hand 30. For example, a closed hand can be associated with a shorter specified time period, and an open hand can be associated with a longer specified time period.

In a first comparison step 54, 3D user interface 20 waits for user 22 to engage virtual keyboard 40. If 3D user interface 20 is engaged, then in a capture step 56, computer 26 captures a sequence of depth maps of a body part such as hand 30.

To engage virtual keyboard 40 (i.e., so that user 22 can input characters via the virtual keyboard), user 22 can move hand 30 so that the 3D user interface presents cursor 42 within the presented virtual keyboard. To disengage from virtual keyboard 40, user 22 can move hand 30 randomly so that 3D user interface 20 does not present cursor 42 in the vicinity of any given virtual key 70 for more than the specified time period. Alternatively, user 22 can disengage from virtual keyboard 40 by moving hand 30 so that the 3D user interface presents cursor 42 outside virtual keyboard 40. In some embodiments, 3D user interface 20 can convey visual feedback when user 20 engages and disengages from virtual keyboard 40. For example, the 3D user interface can change the shading (or color) of virtual keyboard 40 when the user engages and disengages the virtual keyboard.

As discussed supra, user 22 can control cursor 42 by moving hand 30 (or any other limb) in X-Y plane 38, and select a given virtual key 70 by positioning hand 30 so that cursor 42 is positioned in proximity to the given virtual key (i.e., either over the given virtual key or within the border of the given virtual key) for the specified time period. As user 22 moves hand 30 in X-Y plane 38, 3D user interface 20, in a presentation step 58 presents cursor 42 at positions indicated by the hand in the captured sequence of depth maps.

In embodiments of the present invention, computer 26 selects one of virtual keys 70 upon the captured sequence of depth maps indicating an interruption of a motion of cursor 42 (i.e., in response to an interruption of a motion of hand 30 or any other body part) in proximity to the one of the multiple virtual keys. As described in detail hereinbelow, the interruption of the motion may comprise (a) user 30 maintaining hand 30 relatively stationary for either a standard or an override time period as computer 26 presents cursor 42 in proximity to the one of the multiple virtual keys, or (b) user 30 changes direction of hand 30 in proximity to the one of the multiple virtual keys.

In a second comparison step 60, if the captured sequence of depth maps indicate a specified change in direction of a trajectory of hand 30 (i.e., without the hand pausing for at least the standard time period), then in a model application step 62, computer 26 executes a language model that attempts to select one of virtual keys 70 that is in proximity to cursor 42 as the cursor changes direction. However, if the captured sequence of depth maps does not indicate a specified change in direction of a trajectory of hand 30, then the method continues with a third comparison step 64.

In the third comparison step, if user 22 keeps hand 30 relatively steady so that computer 26 presents cursor 42 in proximity to a given virtual key 70 (i.e., within border 72, or adjacent to the given virtual key) for the standard time period (e.g., 0.1 seconds), then the method continues with step 62, where the language model checks if a character associated with the given virtual key comprises a character predicted by the language model. However, if user 22 moves hand 30 so that computer 26 does not present cursor 42 in proximity to a given key 70 for the standard time period (i.e., less than the standard time period), then the method continues with step 54.

Typically, the language model executed in step 62 analyzes the virtual keys that are in proximity to cursor 42 as the cursor changes direction, and selects one or more virtual keys 70 that best appends to any text (i.e., a sequence of one or more virtual keys 70) previously selected and presented in text box 74. Note that there may be instances when the language model does not select any virtual key 70, if none of the virtual keys that are in proximity to cursor 42 as the cursor changes direction are sufficiently probable.

In some embodiments, the language model may apply rules specific to a given language (e.g., English), including but not limited to word rules, short phrase rules, parts of speech rules and grammatical rules. In additional embodiments the language model may utilize information on user 22 who is interacting with the virtual keyboard, including but not limited to a custom dictionary based on text previously entered by the user during a related input session (i.e., text input via the virtual keyboard or any other input device).

For example, if user 22 previously entered the words “Mozart” and “Beethoven” via virtual 40, the language model may set a parameter that indicates that the user prefers classical music. Therefore, if the user enters the word “Bavj” via the virtual keyboard, the language model may correct “Bavj” to “Bach” (“v” is adjacent to “c” and “j” is adjacent” and “h” on the virtual keyboard), even though “Bach” was not explicitly added to the dictionary during previous input session to the music selection field. Note that “navy” is another interpretation for similar motion (with a single key shift relative to the intended “Bach”), but will be less favorable by the language model, as previous text associated with classical music was already entered by the user.

In further embodiments, the language model may utilize an expected semantic domain. For example, the language model may select a response using a dictionary custom tailored to a question or a field type that 3D user interface 20 presents on display 28. In other words, the language model may utilize a custom dictionary specific to an application executing on computer 26. For example, if 3D user interface 20 presents an input field on display 28 for a movie title or a book title, the language model can utilize a dictionary of movie and/or book titles. As an additional example, if computer 26 is executing an adventure-type game, the language model can look for specific commands (e.g., RUN, STOP, FIRE, HIDE, etc.). As a further example, if 3D user interface is presenting a personal information form to be filled out by user 22, the language model can look for specific values for each field (e.g., “M” or “F” for the user's sex).

Examples of language models that can be implemented by computer 26 include a dictionary and statistical models including but not limited to a statistical dictionary, an n-gram model, a Markov model, and a dynamic Bayesian network. Language models are described in further detail in the book “Foundations of Statistical Natural Language Processing”, by Christopher D. Manning and Hinrich Schütze, MIT Press, 1999, Chapters 6, 7, 9 and 12, which is incorporated herein by reference.

In a fourth comparison step 66, if computer 26 selects one or more relevant (i.e., to the language model) virtual keys 70 that are in proximity to cursor 42 (i.e., as the cursor either changes direction or is in proximity to the given virtual key for the standard time period), then the computer presents the one or more selected virtual keys in text box 74 as visual feedback in a presentation step 67, and the method continues with step 54. However, if computer 26 (i.e., since the language model did not select any of the virtual keys) does not select any virtual key 70 in the fourth comparison step, then the method continues with a fifth comparison step 68.

In the fifth comparison step, if user 22 keeps hand 30 relatively steady so that computer 26 presents cursor 42 in proximity to the given virtual key 70 (i.e., within border 72, or adjacent to the given virtual key) for the override time period (e.g., 0.5 seconds), then the computer selects the given virtual key in a selection step 69, and the method continues with step 67. However, if user 22 moves hand 30 so that computer 26 does not present cursor 42 in proximity to the given key 70 for the override time period, then the method continues with step 54.

In some embodiments, 3D user interface 20 can convey visual feedback to user 22 while selecting a given virtual key 70. For example, the 3D user interface can gradually change the shading (e.g., a gray level) of the given character presented on the given virtual key as user 22 maintains the cursor over the given virtual key. The 3D user interface can accept the given virtual key as an input when the shading reaches a certain level. Additionally or alternatively, 3D user interface 20 can increase the size of the given virtual key after the specified time period, thereby conveying an indication that the given virtual key is being “pressed”.

In additional embodiments, user 22 can repeat the input of a given virtual key 70 twice (e.g., “tt”) by keeping hand 30 relatively stationary so that the 3D user interface 20 maintains the cursor's position over the given virtual key twice as long as the relevant time period (i.e., either the standard or the override time periods). In a similar fashion, user 22 can repeat the input of the given virtual key three or more times. In alternative embodiments, 3D user interface 20 can limit the input of the given virtual key to a single character, regardless of how long cursor 42 is positioned over the given virtual key. To repeat the given virtual key, user 22 moves hand 30 to first position the cursor outside the border of the given virtual key, and then moves the hand a second time to position the cursor back within the border of the given virtual key.

In further embodiments, 3D user interface 20 can be configured to accelerate the rate of virtual keyboard 40 input by monitoring both hands 30 of user 22. The 3D user interface can measure separate distances between each hand 30 and 3D sensing device 24, and identify the hand closer to the 3D sensing device as active, and identify the other hand as inactive. Therefore, while “pressing” a given virtual key 70 with the active hand, the user can position the inactive hand above the next virtual key 70 that the user intends to “press”.

When monitoring both hands of user 22, 3D user interface 20 may present either one or two cursors 42. When presenting a single cursor 42, 3D user interface 20 can toggle the cursor between the active and the inactive hand. In other words, 3D user interface 20 can first position cursor 42 in response to a position of the active hand. Once user 22 has selected a given virtual key 70 with the active hand, 3D user interface 20 can then position cursor 42 in response to a position of the inactive hand. When presenting two cursors 42, user interface 20 may position a first cursor 42 in response to a position of the active hand, and position a second cursor 42 in response to a position of the inactive hand.

FIG. 4 shows cursor 42 traversing path segments 80, 82 and 84 in response to user 22 moving hand 30 in X-Y plane 38, as the user enters the word “no” via virtual keyboard 40. Initially, computer 26 positions cursor 42 over the virtual keys “q” and “w”. User 22 inputs the letter “n” by moving hand 30 in X-Y plane 38, so that 3D user interface 20 moves cursor 42 along path segment 80 to a position over the “n” virtual key. As user 22 keeps hand 30 relatively steady over the “n” virtual key for the standard time period, the 3D user interface accepts “n” as an input, and presents “n” in text box 74.

User 22 then inputs the letter “o” by moving hand 30 in X-Y plane 38, so that 3D user interface 20 moves cursor 42 along path segment 82 to a position over the “o” virtual key. As user 22 keeps hand 30 relatively steady over the “o” virtual key for the standard time period, the 3D user interface accepts “o” as an input and presents “o” in text box 74. Finally, user 22 presses the Enter virtual key by moving hand 30 in X-Y plane 38, so that 3D user interface 20 moves cursor 42 along path segment 84 to a position over the Enter virtual key. As user 22 keeps hand 30 relatively steady over the Enter virtual key for the standard time period, the 3D user interface accepts the Enter key as an input. Note that the example described in FIG. 4 uses the standard time period, since the word “no” is a word that can be predicted by a language model.

FIG. 5 shows cursor 42 traversing path segments 90, 92 and 94 in response to user 22 moving hand 30 in X-Y plane 38, as the user enters the word “not” via virtual keyboard 40. Initially, 3D user interface 20 positions cursor 42 over the virtual keys “q” and “w”. User 22 inputs the letter “n” by moving hand 30 in X-Y plane 38, so that 3D user interface 20 moves cursor 42 along path segment 90 to a position over the “n” virtual key. As user 22 keeps hand 30 relatively steady over the “n” virtual key for the specified time period, the 3D user interface accepts “n” as an input, and presents “n” in text box 74.

User 22 then inputs the letter “o” by moving hand 30 in X-Y plane 38, so that 3D user interface 20 moves cursor 42 along path segment 92 to a position over the “o” virtual key. As user 22 keeps hand 30 relatively steady over the “o” virtual key for the specified time period, the 3D user interface accepts “o” as an input and presents “n” in text box 74.

After entering the letters “n” and “o”, user 22 moves hand 30 in X-Y plane 38 so that 3D user interface 20 moves cursor 42 along a path 94 in response to the hand's movement. Using the captured sequence of depth maps, computer 26 calculates a point 96 along path segment 94, which indicates a change in direction of a trajectory of the cursor, as the cursor crosses over the virtual keys “t”, “r”, “f” and “c”. Computer 26 applies a language model to resolve the ambiguity of multiple possible letters and selects the most likely virtual key 70 that user 22 intended to “press”. In the example shown in FIG. 5, the language model evaluates the probability of the strings “not”, “nor”, “nof” and “noc”, and identifies “not” as the most probable text string.

It will be appreciated that the embodiments described above are cited by way of example, and that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and subcombinations of the various features described hereinabove, as well as variations and modifications thereof which would occur to persons skilled in the art upon reading the foregoing description and which are not disclosed in the prior art.

Claims

1. A method, comprising: presenting, by a computer system executing a non-tactile three dimensional user interface, a virtual keyboard on a display, the virtual keyboard comprising multiple virtual keys;capturing a sequence of depth maps, via a 3D capturing device, over time of a hand of a human subject while the human subject moves the hand in a plane;presenting, on the display, a cursor at positions indicated by the hand in the captured sequence of depth maps such that the cursor moves over the virtual keys on the display in response to movement of the hand in the plane; andselecting one of the multiple virtual keys if a change in direction of trajectory is determined by using the captured sequence of depth maps and calculating points along the presented cursor's trajectory path segment over the virtual keys, wherein the trajectory of the user's hand tracks a plurality of keys for which a cursor crosses spatially from a previously determined key point to a newly desired key point, and wherein the keys between the previously determined key point to the newly desired key point are used in combination with previously determined key inputs to configure a language model for estimating probable words and estimating a most likely key from the keys based on the language model and displaying the most likely key appended with the previously determined key inputs; and wherein if the change of trajectory is not detected then determining whether the cursor is in proximity to a plurality of keys for a standard time period and selecting keys to be used by the language model for estimating a most likely key from the plurality of keys and displaying the most likely key appended with the previously determined key inputs.

2. The method according to claim 1, wherein selecting one of the multiple virtual keys comprises using a language model.

3. The method according to claim 2, wherein the language model is selected from a list consisting of a dictionary, a statistical dictionary, an n-gram model, a Markov model and a dynamic Bayesian network.

4. The method according to claim 2, wherein the language model applies rules specific to a given language.

5. The method according to claim 4, wherein the rules are selected from a list consisting of word rules, short phrase rules, parts of speech rules and grammatical rules.

6. The method according to claim 2, wherein the language model utilizes a custom dictionary based on text previously entered by a user interacting with the non-tactile three dimensional user interface.

7. The method according to claim 2, wherein the language model utilizes a custom dictionary specific to an application executing on the computer system.

8. The method according to claim 2, and comprising the language model selecting none of the virtual keys if none of the virtual keys that are in proximity to the cursor are sufficiently probable.

9. The method according to claim 1, and comprising presenting visual feedback of the selected one of the multiple virtual keys on the display.

10. The method according to claim 1, wherein each of the multiple virtual keys is selected from a list consisting of alphanumeric characters, symbol characters, punctuation characters and control commands.

11. The method according to claim 1, wherein presenting the cursor in proximity to the one of the multiple virtual keys comprises presenting the cursor within a border surrounding the one of the multiple virtual keys.

12. The method according to claim 1, and comprising conveying visual feedback upon presenting the cursor in proximity to the one of the multiple virtual keys.

13. The method according to claim 1, and comprising conveying visual feedback when a user engages and disengages from the virtual keyboard.

14. An apparatus, comprising: a display; anda computer executing a non-tactile three dimensional user interface and configured to present a virtual keyboard on a display, the virtual keyboard comprising multiple virtual keys, to capture a sequence of depth maps, via a 3D capturing device, over time of a hand of a human subject while the human subject moves the hand in a plane, to present, on the display, a cursor at positions indicated by the hand in the captured sequence of depth maps such that the cursor moves over the virtual keys on the display in response to movement of the hand in the plane, and to select one of the multiple virtual keys if a change in direction of trajectory is determined by using the captured sequence of depth maps and calculating points along the presented cursor's trajectory path segment over the virtual keys, wherein the trajectory of the user's hand tracks a plurality of keys for which a cursor crosses spatially from a previously determined key point to a newly desired key point, and wherein the keys between the previously determined key point to the newly desired key point are used in combination with previously determined key inputs to configure a language model for estimating probable words and estimating a most likely key from the keys based on the language model and displaying the most likely key appended with the previously determined key inputs; and wherein if the change of trajectory is not detected then determining whether the cursor is in proximity to a plurality of keys for a standard time period and selecting keys to be used by the language model for estimating a most likely key from the plurality of keys and displaying the most likely key appended with the previously determined key inputs.

15. The apparatus according to claim 14, wherein the computer is configured to select one of the multiple virtual keys by using a language model.

16. The apparatus according to claim 15, wherein the computer is configured to select the language model from a list consisting of a dictionary, a statistical dictionary, an n-gram model, a Markov model and a dynamic Bayesian network.

17. The apparatus according to claim 15, wherein the language model applies rules specific to a given language.

18. The apparatus according to claim 17, wherein the computer is configured to select the rules from a list consisting of word rules, short phrase rules, parts of speech rules and grammatical rules.

19. The apparatus according to claim 15, wherein the language model utilizes a custom dictionary based on text previously entered by a user interacting with the non-tactile three dimensional user interface.

20. The apparatus according to claim 15, wherein the language model utilizes a custom dictionary specific to an application executing on the computer system.

21. The apparatus according to claim 15, wherein the computer is configured to select none of the virtual keys if none of the virtual keys that are in proximity to the cursor are sufficiently probable.

22. The apparatus according to claim 14, wherein the computer is configured to present visual feedback of the selected one of the multiple virtual keys on the display.

23. The apparatus according to claim 14, wherein the computer is configured to select each of the multiple virtual keys from a list consisting of alphanumeric characters, symbol characters, punctuation characters and control commands.

24. The apparatus according to claim 14, wherein the computer is configured to present the cursor in proximity to the one of the multiple virtual keys by presenting the cursor within a border surrounding the one of the multiple virtual keys.

25. The apparatus according to claim 14, wherein the computer is configured to conveying visual feedback upon presenting the cursor in proximity to the one of the multiple virtual keys.

26. The apparatus according to claim 14, wherein the computer is configured to conveying visual feedback when a user engages and disengages from the virtual keyboard.

27. A computer software product comprising a non-transitory computer-readable medium, in which program instructions are stored, which instructions, when read by a computer executing a non-tactile three dimensional user interface, cause the computer: to present a virtual keyboard on a display, the virtual keyboard comprising multiple virtual keys, to capture a sequence of depth maps, via a 3D capturing device, over time of a hand of a human subject while the human subject moves the hand in a plane, to present on the display, a cursor at positions indicated by the hand in the captured sequence of depth maps such that the cursor moves over the virtual keys on the display in response to movement of the hand in the plane, and to select one of the multiple virtual keys if a change in direction of trajectory is determined by using the captured sequence of depth maps and calculating points along the presented cursor's trajectory path segment over the virtual keys, wherein the trajectory of the user's hand tracks a plurality of keys for which a cursor crosses spatially from a previously determined key point to a newly desired key point, and wherein the keys between the previously determined key point to the newly desired key point are used in combination with previously determined key inputs to configure a language model for estimating probable words and estimating a most likely key from the keys based on the language model and displaying the most likely key appended with the previously determined key inputs; and wherein if the change of trajectory is not detected then determining whether the cursor is in proximity to a plurality of keys for a standard time period and selecting keys to be used by the language model for estimating a most likely key from the plurality of keys and displaying the most likely key appended with the previously determined key inputs.