WO2010022866A2

WO2010022866A2 - Translation apparatus

Info

Publication number: WO2010022866A2
Application number: PCT/EP2009/005924
Authority: WO
Inventors: Cortland Cecil Williamson
Original assignee: Health Enterprise East Ltd
Current assignee: Health Enterprise East Ltd
Priority date: 2008-08-28
Filing date: 2009-08-14
Publication date: 2010-03-04
Anticipated expiration: 2011-02-28
Also published as: WO2010022866A3

Abstract

A translation apparatus, comprising a processor; at least one input device; at least one output device; and a data storage device; wherein the data storage device comprises a phrase database, the phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; wherein the phrase database further comprises a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; the processor comprising a output interface connected to the at least one output device, an input interface connected to the at least one input device and a data storage interface connected to the data storage device, wherein the processor is programmed so as to be operable to: allow selection of a phrase using the input device so as to provide a selected phrase; retrieve the question and response entry for the selected phrase from the phrase database; and cause the at least one output device to output the selected phrase in both the first language and the second language together with the plurality of responses to the selected phrase in the second language.

Description

TITLE OF THE INVENTION

TRANSLATION APPARATUS

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application No. 61/092,724, filed August 28, 2008.

TECHNICAL FIELD

This invention relates to a translation apparatus, and to a computer- implemented method of translation.

BACKGROUND

In today's metropolitan environments where many cultures and languages co-exist, medical personnel face communication and translation challenges in emergency and non-emergency situations where medical care is required. There is a lack of immediate availability of interpreting services, particularly at General Practitioner, Family Physician or Dentist offices as well as in ambulances and hospital Emergency Rooms. Access to traditional interpreters or telephone translation lines requires time, effort and cost, which can result in delays in treating the patient and potential for incorrect treatment. In settings such as ambulances or emergency care, there is simply no time to access an interpreter, whether in person or by telephone.

The use of dictionaries or phrasebooks, an alternative to interpreters, by non-linguistically trained staff is slow and fraught with errors.

Furthermore, in order to be of use, every medical practitioner would have to carry a phrase book covering every language they were likely to encounter, and then find the correct language and the correct phrase at speed in a potentially fast moving situation. This is unsatisfactory.

BRIEF SUMMARY OF THE INVENTION

According to an aspect of the invention, there is provided a translation apparatus, comprising: a processor; at least one input device; at least one output device; and a data storage device; wherein the data storage device comprises a phrase database, the phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; wherein the phrase database further comprises a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; the processor comprising a output interface connected to the at least one output device, an input interface connected to the at least one input device and a data storage interface connected to the data storage device, wherein the processor is programmed so as to be operable to: i) allow selection of a phrase using the input device so as to provide a selected phrase; retrieve the question and response entry for the selected phrase from the phrase database in response to the selection of a phrase; and iv) cause the at least one output device to output the selected phrase in both the first language and the second language together with the plurality of responses to the phrase selected in the second language.

In an alternative aspect of the invention, the invention comprises a system for multilingual medical information interpretation comprising a laptop, desktop or mobile computing environment and a real time multilingual web based medical interpretation computer program providing translation between a patient and English speaking (or local/natural language speaking) medical personnel. The system may be web based, such that patient responses can be accessed by technicians, nurses and doctors throughout every stage of the patient's care.

The desktop software proposed by this aspect of the invention for medical translation needs can be a system providing speech recognition on the English (or local, natural, language) side of the conversation (with backup of keyboard/mouse interface) to select phrases, translate according to phrasebook and use voice tags to communicate the phrase in the foreign (or second) language to the patient. Further, the patient will then use the keyboard/mouse interface to select the response from among the likely responses in the foreign (or second) language, which is translated back into English (or local, natural language) for the Physician or other user. The system will then build on this simple mechanism to allow wider phrase selection and composite phrase selection, adaptive learning on the types of questions each physician uses to refine the system for better dialog with non-English (non-local) speaking patient population. The system will incorporate tracking, auditing features to enable further learning and development.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING(S)

Figure 1 shows a schematic view of a translation apparatus according to a first embodiment of the invention;

Figure 2 shows the contents of the data storage device of Figure 1 ;

Figure 3 shows entries in the phrase database of Figure 2;

Figure 4 shows a flowchart depicting the method of operation of the apparatus of Figure 1 in a first mode; Figure 5 shows a sample display to the users of the method of Figure 4;

Figure 6 shows a flowchart depicting the method of operation of the apparatus of Figure 2 in a second mode;

Figure 7 shows a sample display to the users of the method of Figure 6; and

Figure 8 shows a translation apparatus according to a second embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

A translation apparatus according to a first embodiment of the invention is shown in Figure 1 of the accompanying drawings. It comprises a general purpose personal computer 1 , running an operating system such as Microsoft Windows, Linux or Apple Mac OS X.

The computer 1 comprises a processor 2 and a data storage device 3 coupled thereto. The data storage device 3 will typically comprise a hard disk, but could also comprise random access memory (RAM), particularly flash RAM, or any other readable, and preferably recordable, storage. The data storage device is coupled to the processor by means of a data storage interface 8 (such as a Serial ATA interface).

The apparatus also comprises input devices of the form of a keyboard 4 and mouse 5, and output devices of the form of a loudspeaker 6 and a display 7. The input and output devices are coupled to the processor by means of input 9 and output 10 interfaces, such as USB interfaces or sound or video cards. Turning now to Figure 2, this shows some of the contents of the data storage device 3. The data storage device stores programming 10 for the personal computer 1 , which causes the processor to act in the manner described below.

The data storage device also stores a phrase database 1 1. This phrase database 1 1 is shown in more detail in Figure 3. The phrase database comprises a plurality of question and response entries; each question and response entry 12 comprises a phrase, usually a question, in a first language, in this example English, and the same phrase in a second language, in this example Kurdish. Each entry also comprises a link to a plurality of responses 13 to the phrase in the second language.

The responses are stored as a plurality of response entries 13 in the phrase database, which each comprise the response in the second language and the first language. For example, a question and response entry 12 could indicate the English and Kurdish versions of the question "Where does it hurt?", and link to Kurdish and English versions of the responses 13 "my leg", "my head", "my abdomen" and so on.

Returning to Figure 2, the data storage device 3 also stores a profile database 21. This comprises a series of user profiles. Each user registered with the translation apparatus can be allocated one or more profiles; in the example of a doctor's surgery, the profiles could be doctor, nurse, receptionist, telephonist, administrator or so on. Each profile allows access to a certain set of the phrases in the phrase database 1 1 ; for example, a receptionist may be allowed access only to those phrases which would allow them to book a patient an appointment, whereas a doctor would have access to a full set of medical phrases and responses. The data storage device 3 also stores a conversation database 31. This records the details of the exchanges between users as set out in the example below.

This phrase database is used as shown in the flowchart in Figure 4. In this example, two users are using the apparatus of Figure 1. User A is a doctor (and so has a doctor's profile, allowing him unrestricted access to the phrase database), seeking information about a patient, user B. User A understands only English, whilst user B understands only Kurdish.

In step 100, the translation apparatus displays on the display 7 a list of phrases that can be selected by a user. Initially, user A may be provided with a list of all the phrases in the phrase database. However, they can instead be provided with a set of pre-defined conversations, comprising a list of questions to be asked in order. Conversations could include standard medical questions or protocols, standard administrative questions or similar.

Conditional branching logic can be built in to the conversation, so that certain questions are only asked if previous questions are answered in a certain manner. User A then selects the phrase directly or by choosing a conversation using the keyboard 4 or mouse 5.

In step 102, the processor retrieves the question and response entry for the phrase selected in step 100, and displays that on the display. Such a display is shown in Figure 5. The screen displayed to the users comprises a header 49 identifying the patient in question.

In this example, user A has selected the phrase "What is your immigration status?", presumably so that user B's entitlement to health benefits can be assessed. The processor has retrieved the question and response entry from the database, and so displays this phrase in English at 50 and in Kurdish at 51. The question and response entry 12 may include an audible rendition of the phrase; this can be played through loudspeaker 6 at this point.

The processor also retrieves the response entries 13 referred to by the question and response entry 12 in question. The Kurdish responses are displayed at 52a, 52b, 52c, along with their English translations 53a, 53b,

53c - here "Asylum seeker", "Failed Asylum seeker" and "Refugee".

Again, each of these can have an audible rendition associated with them in the phrase database, which can be played to user B in the case that they cannot read.

Some of the questions or the responses may be partial phrases, where further information is required in order to complete the phrase. For example, User A may want to tell User B to take these pills X times per day; there may therefore be provided in the phrase database a phrase saying "Take these X times per day". Before choosing the phrase, User A will complete X, for example by picking "three". User A will be provided with sub-phrases for the further information, which will be stored in the phrase database along with their translations into Kurdish. Once user A completes the phrase and selects it, User B will see the phrase "Take these three times per day" in Kurdish.

Similarly, user A may ask user B a question "for how long have you had these symptoms?". User B may be provided with a response "For X", where X can be selected from the group of phrases "a week", "a fortnight", "a month" or so on. These sub-phrases will be stored in the phrase database in both English and Kurdish, so that user B can select them in Kurdish, but the whole response can be displayed in English to user A at step 106 discussed below.

At step 104, User B indicates which of the entries is relevant to them, and either selects the correct response themselves with the keyboard 4 or mouse 5, or asks user A to do so. In the alternative, a free text field is provided 54 in which user B can record their response in their own language (here Kurdish) for later translation by a competent translator, or attempt an English entry if they are feeling sufficiently conversant.

At step, 106, the translation of the response into English is displayed for User A. Accordingly, so far a question has been posed and answer made. These steps can be repeated as many times as are necessary to complete a conversation, be it to diagnose a health problem or to collect necessary administrative information.

Each question and response can be recorded in the conversation database. Each conversation - that is a series of questions and responses - between a unique pair of users can be recorded separately. The conversations of a particular user - say, a patient - can be easily be accessed or made part of their medical records, and can be rendered in the appropriate language for the operation keeping the records. In the example of our doctor's surgery, this would be English, despite the patient user B not understanding that language.

Furthermore, a field 55 for notes entered by the doctor (user A) is provided; these will be stored in the conversation database with the conversation, thus providing a medical context to the conversation. For example, a doctor could report that user's B's foot was red and inflamed should they report that their foot hurt.

In an alternative mode depicted in Figure 6, user A (the doctor) can provide a pre-prepared form for user B (the patient) to fill in. In step 150, the doctor selects a series of phrases from the phrase database (such as name, asylum status, length of time in country, symptoms). Each phrase corresponds to a question and response entry in the phrase database. The doctor can therefore generate a standard form in his language, English. In step 152, and as shown in Figure 7, this form can then be displayed to user B using the display 7 (in Figure 7, only a single question is shown in detail, whereas subsequent questions are only shown in outline); by retrieving the appropriate question and response entries, the processor 2 can then cause the display to display the form in user B's language, that is Kurdish. As shown in Figure 7, the associated responses linked from the question and response entry for that phrase are displayed in a drop down box 61 adjacent to the question 60.

In step 154, User B then uses the mouse 4 and keyboard 5 to select a response in Kurdish to each question.

Each response is recorded, and the response entries in the phrase database 1 1 used to convert the responses back to English. In step 156, the completed form can then be displayed to user A (or anyone else thought appropriate) in English, or printed out or stored for recordal.

Whilst the present embodiments have been described with the databases 1 1 , 21 , 31 on the computer 1 at which the users are located, the databases may be stored remotely, with access over the Internet through a web-based portal. An example of such an embodiment is shown in Figure 8. In this, a local terminal 200 is provided, which comprise a computer 201 having the hardware features of the computer 1 of the first embodiment; corresponding reference numerals have been used, raised by 200.

In addition to the hardware of the previous embodiment, the processor is provided with a network adaptor 220, such as an Ethernet adaptor or an ADSL (asymmetric digital subscriber line) modem. This connects to the Internet 250, through the usual intermediary of an Internet Service Provided (ISP). Also connected to the Internet is a server 300. This comprises a computer 301 comprising a processor 302 and a data storage device 303. The databases 1 1 , 21 , 31 are stored on this data storage device 303. The processor 302 is provided with a network adaptor 320, by means of which the server is connected to the Internet 250, again through usual intermediary of an Internet Service Provider (ISP).

In this apparatus, the server 300 provides a web interface to the phrase database 1 1. The embodiment functions in the same way as the first embodiment, apart from the fact that the input from the users at the local terminal 200 is processed by that terminal and converted into hypertext markup language (HTML) requests, and transmitted to the server 300 over the Internet 250 by means of the hypertext transfer protocol (HTTP), preferably over a secure socket layer (SSL), a combination referred to as HTTPS, so as to secure potentially confidential information exchanges. Output to the users is provided in the form of web pages in HTML, transmitted back across the Internet 250 to the local terminal 200, again via HTTPS.

Whilst the Internet 250 is proposed in the present embodiment, any suitable network, such as a secure private network (which would most likely be more secure than the Internet), may be used.

Throughout the above embodiments, it is to be noted that the languages chosen (English and Kurdish) are examples only, and any other pair of languages in which it is desired to communicate could be substituted therefor. Furthermore, whilst the embodiments are discussed with reference to a medical practice, the invention is applicable in any situation where it is desirable to translate a conversation between two users.

Claims

1. A translation apparatus, comprising: a) a processor; b) at least one input device; c) at least one output device; and d) a data storage device; wherein the data storage device comprises a phrase database, the phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; wherein the phrase database further comprises a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; the processor comprising a output interface connected to the at least one output device, an input interface connected to the at least one input device and a data storage interface connected to the data storage device, wherein the processor is programmed so as to be operable to: i) allow selection of a phrase using the input device so as to provide a selected phrase; ii) retrieve the question and response entry for the selected phrase from the phrase database; and iv) cause the at least one output device to output the selected phrase in both the first language and the second language together with the plurality of responses to the selected phrase in the second language.

2. The translation apparatus of claim 1 , wherein the processor is programmed so as to be operable to retrieve the question and response entry for the selected phrase from a group of phrases, based on the question.

3. The translation apparatus of claim 1 , wherein the processor is programmed so as to be operable to retrieve certain questions, only if previous questions are asked in a certain manner.

4. The translation apparatus of any preceding claim, wherein the processor is programmed so as to be operable to allow selection of one of the plurality of responses to the selected phrase in the second language using the input device to provide a selected response and to, in response, cause the at least one output device to output the selected response in the first language.

5. The translation apparatus of any preceding claim, wherein the processor is programmed so as to be operable to cause the at least one output device to output the plurality of responses in the first language together with the plurality of responses in the second language.

6. The translation apparatus of any preceding claim, wherein at least one of the question and response entries is a compound phrase entry, each compound phrase entry comprising a phrase comprising a fixed part and a variable part, and wherein the processor is programmed so as to be operable to allow completion of the variable part of the phrase in the first language using the at least one input device before causing the output device to output the fixed part and the variable part, as completed, of the phrase in the first and second language.

7. The translation apparatus of any preceding claim, wherein the at least one output device comprises at least one of a display and a loudspeaker.

8. The translation apparatus of any preceding claim, wherein the at least one input device comprises at least one of a keyboard and a pointing device.

9. The translation apparatus of any preceding claim, wherein the data storage device comprises a profile database comprising a plurality of profiles, wherein each profile identifies a selection of question and answer entries which are application to a class of user.

10. The translation apparatus of any preceding claim, wherein the data storage device comprises a conversation database, wherein the processor is programmed so as to be operable to record each selection of a phrase and each selection of a response in the conversation database.

1 1. The translation apparatus of any preceding claim, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a plurality of further second languages and indicate a plurality of responses to the phrase in the plurality of further second languages.

12. The translation apparatus of any preceding claim, in which the input interface and the output interface are network interfaces, and the apparatus comprises a network connecting at least one of the at least one input device and the at least one output device to at least one of the processor and the data storage device.

13. The translation apparatus of any preceding claim, in which the network is the Internet.

14. A translation apparatus, comprising: a) a processor; b) at least one input device; c) at least one output device; and d) a data storage device; wherein the data storage device comprises a phrase database, the phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; wherein the phrase database further comprises a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; the processor comprising a output interface connected to the at least one output device, an input interface connected to the at least one input device and a data storage interface connected to the data storage device, wherein the processor is operable to: i) allow a first user to define, using the at least one input device, a series of phrases comprising a plurality of the phrases in the first language, to be put to a second user; ii) retrieve, from the phrase database, the question and response entries relating to each phrase of the series of phrases; iii) output, to the second user using the at least one output device, each phrase of the series of phrases in the second language and the responses, in the second language to each of the phrases of the series of phrases; iv) allow selection of a response for each phrase in the series of phrases by the second user using the at least one input device.

15. The apparatus of claim 14, in which the processor is operable to; v) output, to the first user using the at least one output device, the responses selected by the second user in the first language.

16. A computer-implemented method of translation, comprising the steps of: a) providing stored on a data storage device a phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; b) providing stored in the phrase database a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; c) providing a computer having at least one input device and at least one output device; d) allowing a user to select a phrase using the at least one input device so as to provide a selected phrase, the selection of a phrase causing the computer to retrieve of the question and response entry for the selected phrase from the phrase database; e) outputting, using the at least one output device the selected phrase in both the first language and the second language together with the plurality of responses to the selected phrase in the second language.

17. The method of claim 16, further comprising the step of allowing a user to select of one of the plurality of responses to the selected phrase in the second language using the at least one input device to provide a selected response and, in response, outputting, using the at least one output device, the selected response in the first language.

18. The method of claim 16, further comprising the step of outputting, using the at least one output device, the plurality of responses in the first language together with the plurality of responses in the second language.

19. The method of claim 16, wherein at least one of the question and response entries is a compound phrase entry, each compound phrase entry comprising a phrase comprising a fixed part and a variable part, and wherein step (iv) of the method comprises allowing completion of the variable part of the phrase in the first language using the at least one input device before causing the at least one output device to output the fixed part and the variable part, as completed, of the phrase in the first and second language.

20. The method of claim 16, comprising the step of recording each selection of a phrase and each selection of a response in the conversation database.

21. A computer-implemented method of translation, comprising the steps of: a) providing stored on a data storage device a phrase database comprising a plurality of question and response entries, wherein each of the question and response entries map a phrase in a first language to a translation of the phrase into a second language and indicate a plurality of responses to the phrase in the second language; b) providing stored in the phrase database a plurality of response entries, each of the plurality of response entries mapping one of the plurality of responses in the second language to a translation of the response into the first language; c) providing a computer having at least one input device and at least one output device; d) allowing a first user to define, using the at least one input device, a series of phrases comprising a plurality of the phrases in the first language, to be put to a second user, and storing the series of phrases; e) in response to step (d), causing the computer to retrieve, from the phrase database, the question and response entries relating to each phrase of the series of phrases; f) outputting, to the second user using the at least one output device, the each phrase of the series of phrases in the second language and the responses, in the second language to each of the phrases of the series of phrases; iv) allowing the second user to select a response for each phrase in the series of phrases using the at least one input device.

22. The method of claim 21 , comprising outputting, to the first user using the at least one output device, the responses selected by the second user in the first language.