JP2004206116A - Improved speech recognition user interface for ultrasonic wave system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To disclose an improved ultrasonic wave system which is brought under speech control. <P>SOLUTION: In one aspect, the efficiency and usability of a speech recognition system are improved by using target speech recognition, speech text converting capability, keyword actuation, and user profiles for respective ultrasonic wave system operators. In another aspect, a mechanical user interface (between users and the speech recognition system) is made efficient and simple by using a wireless communication link and a user interface having a headset, a track ball device, and choice input buttons. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

  The present invention relates to an ultrasonic imaging system, and more particularly, to an ultrasonic imaging system having a speech control system.

  Ultrasound systems available today are manually controlled by a keyboard, mouse, trackball, knob, button control, and / or touch panel. It is also possible to add a voice control system to these control means, or to replace these control means with a voice control system. As shown in FIG. 1, such a voice control system in its simplest form includes a microphone 110, a speech recognition system 150, and an ultrasound imaging system 180. The user speaks into microphone 110, which converts speech into electrical signals and sends the electrical signals to speech recognition system 150. The speech recognition system 150 interprets the electric signal, determines what command has been given, and transmits the determined command to the ultrasonic imaging system 180. A specific example of a voice activated (and / or manual) control system is disclosed in US Pat. In U.S. Pat. No. 6,037,095, a speech recognition system 150 is incorporated into an ultrasound system 180.

  While such voice controlled ultrasound systems make it easier for users to do their work, they also have their own problems. First, the user may have to "train" the speech recognition system so that the user can recognize how to pronounce certain words and syllables. In many speech recognition systems, this is accomplished by a "voice registration training session" in which the user repeats certain terms, expressions, and syllables and initializes the system. However, in a hospital or clinic environment, users may use different ultrasound imaging systems and / or different users may use the same ultrasound system. In the former case, the user must perform multiple training sessions (once for each system); in the latter case, the ultrasound system may not have the capacity to store one or more training files.

Second, typically, the microphone is left turned on during and even after use of the system, so that the speech recognition system communicates with the operator (eg, talks to the patient). A large amount of unrelated data is received. This can be misleading by the speech recognition system and result in unintended commands being transmitted to the ultrasound system.
U.S. Pat. No. 5,544,654

  Therefore, there is a need for an improved ultrasound imaging system that reduces misunderstandings by the speech recognition system and increases the efficiency of the user interface.

  The present invention provides improvements in speech controlled ultrasound systems. In one aspect of the invention, the efficiency and ease of use of a speech recognition system is based on training profiles for individual users, targeted speech recognition, speech-to-text conversion capabilities, and keyword activation. And be improved. A user profile is created for each operator, each user profile having at least an operator training profile, a modifiable speech-text dictionary, and a modifiable list of keywords. In another aspect, the mechanical user interface (between the user and the speech recognition system) is streamlined and used by using a wireless communication link and a user interface having a headset, a trackball device, and a selection input button. Be simplified. It is also conceivable to use automatic user identification means, such as voice recognition technology and / or wireless identification transponders, to identify the user in order to retrieve the appropriate user profile.

  Other objects and features of the present invention will become apparent from consideration of the following detailed description in conjunction with the accompanying drawings. It should be understood, however, that the drawings are designed for illustrative purposes and do not define limitations of the invention. Reference should be made to the appended claims for limitations of the invention. Furthermore, it is to be understood that the drawings are not necessarily drawn to scale, and unless otherwise indicated, are intended only to conceptually describe the structures and procedures described herein.

  In general, it is an object of the present invention to personalize a voice controlled ultrasound system and to make the ultrasound system more efficient. In one aspect of the present invention, the efficiency and ease of use of a speech recognition system is enhanced by several improvements. In another aspect, the mechanical user interface (between the user and the speech recognition system) is streamlined and simplified.

The present invention has six essential improvements to the speech recognition user interface for ultrasound systems. That is,
・ User profile ・ Target speech recognition ・ Speech-text conversion ability ・ Keyword activation ・ Wireless communication link ・ Trackball with “select” button
User Profile -As mentioned above, a typical speech recognition system is a speech registration that trains the speech recognition system so that the user can repeat expressions, words, and / or syllables to understand what the user is talking about. Initialized via session. As further noted above, this can be problematic in a hospital or clinic environment where users use different ultrasound systems and / or different users use the same ultrasound system. In an ultrasound system according to an embodiment of the present invention, training profiles are stored for each individual user, thereby enabling a consistent performance of speech recognition in all ultrasound systems used by the user.

  According to a preferred embodiment of the present invention, a user first conducts a voice registration training session, resulting in a speech training profile being formed. As shown in FIG. 2, the speech training profile 210 is stored in the user profile 200 for the user. This user profile 200 is accessed and utilized each time the particular user activates the speech recognition system of the ultrasound system.

  In order for the system to use the correct user profile, the speech recognition system must first recognize that particular user. This recognition can be made by various means. In one embodiment, a user may key in his identification information directly into a speech recognition system, for example, using a keyboard. Then, the system finds the correct user profile 200 based on the user identification information 205. In another embodiment, a speech recognition system is capable of speech recognition (ie, recognizing and identifying individual speech of a particular user). In this embodiment, the user simply starts speaking and the speech recognition system automatically identifies the user and retrieves the user's profile. In such an embodiment, it is also necessary to store a unique parameter that determines the user's voice, such as the voice identification parameter 215 shown in FIG. In another embodiment, a magnetic strip of a credit card type card, a passive electrical element embedded in a credit card type card that sends a signal when pinged by an active electrical element, an identification number, or Bluetooth or other Other types of user identification information may be used, such as a MAC address transmitted by a wireless device of this type.

  The remaining components shown in the user profile 200 are described below.

  User profiles can be stored in various ways. As shown in the preferred embodiment of FIG. 3, user profiles 310 of users A, B, and C are stored in a portion of the memory of speech recognition system 150. In this embodiment, speech recognition system 150 is incorporated into an ultrasound system. Thus, in the embodiment shown in FIG. 3, one ultrasound system must have enough memory to hold all user profiles. It should be noted that in all embodiments, the speech recognition system may be separate from the ultrasound system, or may be integrated with the ultrasound system.

  As shown in the preferred embodiment of FIG. 4, each ultrasound system 401, 402, and 403 has a communication link 405 leading to a centralized user profile database 410. An exemplary embodiment of this storage means is a hospital, where each ultrasound system is connected to a local area network (LAN). A server having a user profile database is also connected to the LAN. Upon recognizing the user, the speech recognition system requests an appropriate user profile from the user profile database and receives the user profile. In a further preferred embodiment shown in FIG. 5, each user profile 501, 502 and 503 is stored in a small portable memory means. This memory means can transmit the stored user profile to the speech recognition system. As an example, a user may carry a small Sony "Memory Stick" with his user profile. This memory stick can be inserted and read out by reading means 511, 512 and 513 which connect to various speech recognition systems. Any storage means (CD, floppy (registered trademark) disk, embedded semiconductor electronic equipment, etc.) can be used, and any wired or wireless storage readout means can be used.

  As can be seen from the examples described herein, the method of storing user profiles can be from centralized (eg, all user profiles are stored on one server, as in FIG. 4) to distributed (eg, The range can vary between each user carrying their own profile, as in FIG. 5, and embodiments between these extremes (eg, as shown in FIG. 3, each speech recognition system may have several user profiles). Having).

  Further, a combination of centralized and distributed systems can also be used. For example, the user profiles may be stored on an ultrasound system, a networked centralized database, and further on a portable storage unit carried by each user. In such a system, the ultrasound system first checks the identity of the user and then determines whether the user's profile is stored. If not stored, the ultrasound system may download the user's profile directly from the user's storage or from a centralized database over a network. The choice depends on the communication and storage capabilities of the ultrasound system. Such a duplication system is useful when the user profile is lost. In addition, if the centralized database is connected to the Internet, the system would be very beneficial for ultrasound technicians traveling through various facilities.

  Further, it should be understood that the user profile 200 shown in FIG. 2 is a schematic diagram of the possible components in the user profile. In some embodiments, the user profile may include only the user identification information 205 and the user's speech training profile 210. Further, components in the user profile 200 may be stored on various devices. For example, in a system such as that shown in FIG. 4, the voice identification parameters 215 may be stored on a separate ultrasound system and the user profile may be stored on the server 410. In such an embodiment, each speech recognition system first identifies the user using locally stored speech identification parameters, and retrieves the appropriate user profile from server 40.

Target speech recognition— The degree of accuracy of a particular word may be significantly lower than the overall recognition accuracy of the speech recognition system. That is, there may be certain words or expressions that are mistranslated much more frequently than other words or expressions. However, in a typical speech recognition system, the only way to solve such a problem is to repeat the entire initial speech registration training session, which is a problem in which the mistranslation problem may take several words or expressions. Is very inefficient because it can only occur for In one embodiment of the present invention, the user can immediately make a "targeted" correction to the speech recognition system if the speech recognition system mistranslates the spoken word or does not know the spoken word. . Using this feature, shown as target training 211 in the user profile 200 of FIG. 2, the user performs voice registration training directed only to the word or expression in question without repeating the entire initial voice registration session.

Speech-to-text conversion capability- The user of the ultrasound system annotates the images created using at least 5% of the patient's examination time. This annotation is typically used to properly identify the image and has information such as cut planes and anatomy. According to one embodiment of the present invention, the ultrasound system has the ability to convert speech to text without the need to interrupt the exam or to manually enter annotation information. Allows users to annotate images. The speech-to-text conversion module may be implemented by hardware, software, or a combination thereof, and may be integrated into the speech recognition system and / or the ultrasound system, or may be a separate module. To avoid the speech-to-text conversion ability interfering with the speech command function, the speech-to-text conversion ability can be activated by hardware or by a specific speech command. For example, when the user says "start annotation", the speech-to-text conversion capability is activated. The system then waits for a spoken stop expression, timeout, or user interaction with some physical interface to stop annotating.

  Furthermore, the speech-to-text conversion feature of the present invention allows for the addition, deletion, and editing of "dictionaries" of words used by the speech-to-text conversion module to recognize speech. For example, the user may want to delete words that are never used (to avoid mistranslation), or modify the way certain words appear as text in the annotation, ie, abbreviate etc. You may wish that. As shown in FIG. 2 as speech-text dictionary 220 in user profile 200, this module can be continuously updated and modified each time a user accesses his profile.

Keyword activation -As mentioned above, ambient sounds, background conversations, etc., can adversely affect the performance of a speech recognition system, potentially causing the device to operate inadvertently. A speech recognition system according to one embodiment of the present invention minimizes this problem by using a "keyword" before each speech command. Keywords are generally meaningless words that are unlikely to be spoken in everyday conversation. Keyword activation is a better solution than requiring keyboard key presses or an additional confirmation step after each spoken command. In some environments, keywords may not be needed or desired, and in such environments, keyword capabilities are not used.

  FIG. 2 shows one or more keywords 230 as stored in the user profile 200. As with the other modules in FIG. 2, this feature can be continuously updated and modified each time the user accesses his profile.

Wireless Communication-In one preferred embodiment of the present invention, the ultrasound system is implemented using wireless connections between components, rather than wired connections. This wireless connection may use any technology, but preferably uses low-power, short-range wireless technology. Some examples of low-power, short-range wireless technologies include Bluetooth, IEEE 802.11a or 802.11b, HiperLAN, and HomeRF. Magnetic induction, infrared or diffuse infrared technology may also be used. As shown in FIG. 6, the headset 610 has a wireless connection with an ultrasound system 680 (in this example, the speech recognition system is incorporated into the ultrasound system 680). This allows the ultrasound system 680 to have a wireless communication link with a peripheral device, such as a printer 662 or PDA 664, or one or more hospital networks. With this added wireless capability, the ultrasound system 680 can be moved to various locations and only needs to reconnect the power plug to the wall socket. Further, the peripheral components can be easily reconnected to any ultrasound system in the room, so that the peripheral components can be easily moved from one room to another.

  In one preferred embodiment, Bluetooth is used for the wireless communication link. The Bluetooth protocol requires a handshake procedure at the beginning of any communication session between two Bluetooth devices. Thus, Bluetooth can ensure that only one audio input device is registered with the ultrasound system at a time. Bluetooth devices are also short-range, lightweight and have low power consumption. Therefore, Bluetooth devices are ideal as headsets in speech recognition ultrasound systems. Standard RF systems are too long and consume too much power. Infrared systems require a line-of-sight optical path. In addition, Bluetooth transmitters are less likely to interfere with other devices in a hospital or clinic environment that use RF, such as medical telemetry devices.

  In one preferred embodiment, rather than using voice identification to identify individual users, headset 610 wirelessly transmits a numeric code when pinged by a unique identifier (eg, pinged by active electronics). Passive identifier), by means of which the user is identified to the speech recognition system. Thus, when the user is near the ultrasound system, the ultrasound system can automatically identify the user and retrieve the correct user profile. This may be used in the wireless embodiment of FIG. 3 or 4. In another example, the user profile may be embedded or connected to headset 510. In a built-in embodiment, a semiconductor memory chip with a user profile is built into the headset and can be accessed and downloaded by the ultrasound system when a user comes within range of the ultrasound system. This may be the wireless embodiment of FIG. In a connected embodiment, a memory module having a user profile may be connected to headset 610 by wire.

Trackball devices- An additional improvement over conventional ultrasound systems is the simplified mechanical user interface that is incorporated into speech recognition systems. A conventional mechanical user interface for an ultrasound system has a control panel with a keyboard having a plurality of push buttons, toggle switches, rocker switches, rotary knobs, and an integrated trackball. However, conventional user interfaces are relatively large and are located far from the user when the user performs an examination. Such frequent back and forth to the conventional user interface can hinder the progress of the test and can lead to occupational stress such as root canal syndrome. Further, in ultrasound systems that use speech recognition technology for control, a mechanical user interface should be incorporated into and support the speech recognition system to control the ultrasound system. .

  In one preferred embodiment of the present invention, shown in FIG. 7, a simplified mechanical user interface 700 includes a trackball 710, a "select" button 720, and a reconfigurable touch panel 730. In other embodiments, the simplified mechanical user interface may have fewer or more components. For example, in another embodiment, the simplified mechanical user interface may only have a trackball with a built-in selection function (ie, activated by pressing the trackball). In yet another embodiment, a simplified mechanical user interface may have all the components shown in FIG. 7 and additional mechanical controls such as toggle switches or rotary knobs. In yet another embodiment, the reconfigurable touch panel 730 is replaced by standard buttons, switches, and / or other mechanical interfaces. An important aspect of the simplified mechanical user interface according to various embodiments of the present invention is that it is integrated into the speech recognition user interface of the ultrasound system and its size and / or size as compared to conventional mechanical user interfaces. Or, the complexity is reduced.

  In FIG. 7, a trackball 710 and a select button 720 cooperate with a headset (ie, a speech recognition system) to form a primary user interface with an ultrasound system. The control of the system can be activated by a speech command and “send” using the select button 720. Trackball 710 provides a simple means of replicating the functions of paddle switches and rotary switches of a conventional user interface by turning control settings at high speed. In some situations, the speech command may display a menu with possible choices on the screen. The user selects one of the choices by moving the trackball 710, indicating the desired choice, and pressing the select button 720. In other situations, actuation of the trackball 710 and / or pressing of the select button 720 causes the speech recognition system to wait for a particular type of speech command or causes the speech recognition system to Change the displayed content. Any of the three input components and combinations of input sequences and / or inputs via the input means shown on the reconfigurable touch panel 730 can be varied without limitation.

  Command functions that are not suitable for speech commands can be performed by the trackball 710 and / or the reconfigurable touch panel 730 in cooperation with the headset. For example, an operation requiring a subtle physical operation such as positioning of a measurement cursor may be performed by the trackball 710. In a preferred embodiment, the reconfigurable touch panel 730 changes the display ("buttons" and other machine-based interface images) based on the current context provided by the speech recognition system. Thus, the “buttons” and other controls on the reconfigurable touch panel 730 may be completely different depending on what the user is currently doing. Further, it is conceivable that the embodiment of the present invention has a display content determined by the user on the reconfigurable touch panel 730. Further, if the speech recognition system cannot determine the command spoken by the user, the trackball 710, the select button 720, and / or the reconfigurable touch panel 730 provide various means to the user, Allows to select more than one option generated by the speech recognition system.

  As shown in FIG. 8, one preferred embodiment of the present invention has a simplified mechanical user interface 700 that is attached to an articulated support 810 that is attached to an ultrasound system 880. In the simplified mechanical user interface 700 in this embodiment, the select button function is built into the trackball. When the trackball is tapped or pressed down, this is the same as pressing the select button. The articulated support 810 allows the user to move the simplified mechanical user interface 700 to a convenient location. In some embodiments, a plurality of simplified mechanical user interfaces 700 are placed in appropriate locations (eg, next to the ultrasound system, on the user's seat, on the right side of the bed, on the left side of the bed, etc.). From there, the user may select the one that is best suited for a particular operation. In other embodiments, the simplified mechanical user interface 700 has a wired or wireless communication link with the ultrasound system, thereby allowing a greater degree of freedom of operation.

  Thus, while the basic novel features of the invention have been illustrated, described, and pointed out with reference to the preferred embodiment of the invention, it is to be understood that the type and details of the apparatus illustrated and its operation It is to be understood that various omissions, substitutions and changes may be made by those skilled in the art without departing from the spirit of the invention. For example, all combinations of components and / or method steps that perform substantially the same function in substantially the same manner to achieve the same result are expressly intended to be within the scope of the present invention. Further, the structures and / or components and / or method steps shown and / or described in any disclosed form or embodiment of the invention may be subject to any other general design choice. The disclosed or described or proposed forms of the present invention or examples may be incorporated. Therefore, it should be limited only by the appended claims.

FIG. 2 is a block diagram illustrating a conventional ultrasound system using a speech recognition system. FIG. 4 is a schematic diagram illustrating a user profile according to a preferred embodiment of the present invention; FIG. 4 illustrates a method for storing various user profiles in one ultrasound system according to one preferred embodiment of the present invention. FIG. 4 illustrates a method of storing various user profiles in one ultrasound system according to another preferred embodiment of the present invention. FIG. 7 illustrates a method of storing various user profiles in one ultrasound system according to a further preferred embodiment of the present invention. 1 is a block diagram illustrating an ultrasound system using a wireless communication link, according to one preferred embodiment of the present invention. FIG. 4 is a block diagram illustrating a simplified mechanical user interface according to one preferred embodiment of the present invention. FIG. 1 is a block diagram illustrating an ultrasound system using a simplified mechanical user interface according to one preferred embodiment of the present invention.

Explanation of reference numerals

110 Microphone 150 Speech recognition system 180 Ultrasound imaging system 200 User profile 205 User identification information 210 Speech training profile 211 Target training 215 Voice identification information 220 Speech-text conversion dictionary 230 Keywords 410 User profile database 401, 402, 403 Ultrasound system 501 , 502, 503 User profile 511, 512, 513 Reading means 610 Headset 662 Printer 664 PDA
680 Ultrasound System 700 Simplified Mechanical User Interface 710 Trackball 720 "Select" Button 730 Reconfigurable Touch Panel 810 Articulated Support 880 Ultrasound System

Claims (20)

A speech recognition system for controlling the ultrasound imaging system;
At least one user profile, the ultrasound imaging system comprising:
The ultrasound imaging system, wherein the at least one user profile comprises a training profile of each user for the speech recognition system of the ultrasound imaging system. Further comprising means for recognizing each user,
The ultrasound imaging system according to claim 1, wherein after the means for recognizing each user has recognized each user, a user profile of each of the recognized users is retrieved. The means for recognizing each user includes:
A tactile input device that allows a user to input data identifying the user;
A voice recognition system in which voice identification parameters of each user are stored in the at least one user profile;
3. The ultrasound imaging system according to claim 2, further comprising at least one of identification means carried by a user and supplying data for identifying the user. The identification means carried by the user,
Memory means connected to the transmitter;
Having at least one of semiconductor memory means readable by a reading device connected to the ultrasonic imaging system,
4. The ultrasound imaging system according to claim 3, wherein the ultrasound imaging system has a corresponding receiver for receiving data identifying the user.   The ultrasound imaging system according to claim 4, wherein the transmitter is a wireless transmitter, and the receiver is a wireless receiver.   The wireless communication link between the wireless transmitter and the wireless receiver is one of a Bluetooth link, an IEEE 802.11 link, a HiperLAN link, a HomeRF link, an infrared technology link, and a magnetic induction technology link. Item 7. An ultrasonic imaging system according to Item 5.   2. The ultrasound imaging system according to claim 1, wherein the speech recognition system includes a unit for performing a target correction on the speech training profile.   2. The ultrasound imaging system according to claim 1, further comprising means for converting spoken speech to text. Further comprising means for annotating the image,
9. The ultrasound imaging system of claim 8, wherein the user annotates by speaking using a means for converting the speech to text.   9. The ultrasound imaging system of claim 8, wherein the at least one user profile comprises a user-modifiable speech-text dictionary. The at least one user profile has at least one keyword;
The ultrasound imaging system of claim 1, wherein the at least one keyword is used by each of the users before speaking a command to the ultrasound imaging system. The ultrasound imaging system is one of a plurality of ultrasound imaging systems,
The at least one user profile comprises:
Memory means in at least one of the plurality of ultrasound imaging systems;
A centralized database that communicates with a plurality of the plurality of ultrasound imaging systems;
The ultrasound imaging system of claim 1, wherein the ultrasound imaging system is located on at least one of a memory means carried by each user and readable by at least one of the plurality of ultrasound imaging systems. Further comprising a user headset having a microphone;
The ultrasound imaging system of claim 1, wherein the headset is in wireless communication with at least the speech recognition system. The ultrasound imaging system is implemented using a wireless communication link between the components,
The ultrasound imaging system of claim 1, wherein the components include a printer, a personal digital assistant (PDA), a user headset, and a user trackball device.   13. The ultrasound imaging system of claim 12, wherein the wireless communication link is one of a Bluetooth link, an IEEE 802.11 link, a HiperLAN link, a HomeRF link, an infrared technology link, and a magnetic induction technology link. Further comprising a simplified mechanical user interface incorporated into the speech recognition system;
The simplified mechanical user interface comprises:
A trackball device for the operator to operate items displayed on the display,
A selection input for an operator to select an item displayed on the display,
The ultrasonic imaging system according to claim 1, further comprising: a touch panel capable of displaying an item and receiving an input. A speech recognition system for controlling the ultrasound imaging system;
Speech recognition training file for each operator, voice recognition system used to identify the operator of the ultrasound imaging system for one of the data entry inputs and to find a corresponding portable speech recognition training file, operator Is a speech-to-text converter for annotating ultrasound images, a speech-to-text dictionary for each operator, modifiable by each operator, and an operator for performing targeted speech recognition training on specific words and expressions. An ultrasound imaging system comprising: means and at least one of a keyword activation feature in which keywords are used by an operator to indicate to the speech recognition system that a spoken command will follow. The ultrasound imaging system is one of a plurality of ultrasound imaging systems,
18. The ultrasound imaging system of claim 17, wherein at least one of the speech recognition training file, the speech-text dictionary, and stored keywords can be carried to any of the plurality of ultrasound imaging systems. . Further comprising a user headset having a microphone;
18. The ultrasound imaging system of claim 17, wherein the headset is in wireless communication with at least the speech recognition system. Display and
A simplified mechanical user interface built into the speech recognition system;
The simplified mechanical user interface comprises:
A trackball device for an operator to operate items displayed on the display,
A selection input for an operator to select an item displayed on the display,
The ultrasonic imaging system according to claim 17, further comprising: a touch panel capable of displaying an item and receiving an input.

Images