WO2014082654A1

WO2014082654A1 - Systems and methods for aiding quran recitation

Info

Publication number: WO2014082654A1
Application number: PCT/EP2012/073682
Authority: WO
Inventors: Walid Magdy; Ahmed M. ALI
Original assignee: Qatar Foundation; Hoarton, Lloyd
Priority date: 2012-11-27
Filing date: 2012-11-27
Publication date: 2014-06-05

Abstract

A mobile smart device (1) incorporates a screen (2) and an audio input (5). The device (1) is operable to display words from the Quran on the screen (2). The device (1) incorporates a speech processing module (6) which is operable to recognise words in an audio signal captured by the audio input (5). The device (1) incorporates a processing arrangement (7) which is operable to recognise words in the captured audio and align the words with words of the Quran stored in a memory (8). The device (1) highlights words of the Quran as they are spoken in substantial real-time on the screen (2). The device (1) highlights incorrectly recited words in a different manner to alert a user to incorrect recitation of the Quran.

Description

Title: Systems and Methods for Aiding Quran Recitation The present invention relates to systems and methods for aiding Quran recitation and memorisation, and more particularly relates to systems and methods for aiding the recitation of the Quran by an Imam during Muslim prayer. In group Muslim prayer, Muslims pray by following an Imam who leads the prayer. In prayers, such as taraweeh during the month of Ramadan, an Imam recites the Quran for long periods of time. Many Muslims prefer to follow the recitation of the Quran by the Imam either by reading from a printed paper version of the Quran (Moshaf) or by reading the Quran on the screen of a mobile smart device, such as a smartphone or tablet. There are three problems that a Muslim may experience when trying to follow a written version of the Quran as it is recited by an Imam:

1 . It can be difficult for Muslims to identify quickly the page and verse of the Quran that is being recited by the Imam. This is because most Muslims do not memorise the full Quran which comprises more than 600 pages and more than 6,000 verses.

2. A Musl im must fl ip pages of the Quran during the recitation if the Musl im is using either a printed version of the Quran or a smart device displaying pages of the Quran. Many scholars find the flipping of pages during prayer to be distracting. Furthermore, it is easy for Muslims to lose their place in the Quran when flipping pages and it can be difficult for a Muslim to find their place again.

3. Non-Arabic speakers may find it especially difficult to align the written Quran with the Imam's recitation. Often the Imam recites the Quran from memory. This can lead to an Imam making mistakes with the recitation of the Quran, such as omitting parts, swapping verses, starting verses from the incorrect place, etc. If an Imam makes a mistake during his recitation, the prayers following the Imam are expected to speak the correct verse out loudly to prompt the Imam to correct the mistake. Occasionally, none of the prayers have memorised the verse which is being recited and so none of the prayers are able to speak out to correct the Imam. This leads to discomfort for both the Imam and the prayers. One solution to this problem would be for the Imam to hold a copy of the Quran either in printed form or on a smart device. However, this solution is not acceptable for the majority of Muslims who expect an Imam to recite the Quran from memory. Indeed, the Imam selection process is influenced by how much of the Quran has been memorised.

Aside from group prayers, Muslims should memorise some parts from the Quran and potentially the full Quran. For a Muslim to test his/her memorisation, he/she tries to recite verses of the Quran from memory with someone who has already memorised the verses or who is holding a Quran book (Moshaf) to alert the reciter if the reciter makes a mistake. Another person is not always available to test a Muslim's memorisation and so many Muslims carry a Moshaf to test themself. A Muslim will recite passages of the Quran without looking at the Moshaf and then check their performance every few seconds to be sure that they are reciting correctly. However, this method can lead to many mistakes in recitation if the Muslim is not paying close attention.

The present invention seeks to provide improved systems and methods for aiding Quran recitation. According to one aspect of the present invention, there is provided a Quran recitation aid comprising: a screen for displaying text from the Quran, an audio input operable to receive an audio signal of a voice reciting text from the Quran, a speech recognition module operable to recognise words in the received audio signal, and a processing arrangement operable to identify the position of the recognised words in the text of the Quran and to provide an indication of the recognised words on the screen.

Preferably, the aid is operable to indicate the recognised words on the screen by highlighting the words in the text from the Quran which is displayed on the screen.

Conveniently, the processing arrangement is operable to provide an alert signal if the recognised words do not appear in the text of the Quran.

Advantageously, the processing arrangement is operable to provide an alert signal if the order of the words recognised by the speech recognition module do not follow the order in which the words appear in the text of the Quran. Preferably, the processing arrangement provides the alert signal by highlighting the words of the Quran that should have been spoken differently in the text of the Quran which is displayed on the screen.

Conveniently, the recitation aid is operable to store a data file representing the history of the accuracy of the Quran recitation.

Advantageously, the recitation aid is operable to display the history of the accuracy of a recitation by highlighting the correctly recited words in the text of the Quran on the screen and highlighting the incorrectly recited words differently in the text of the Quran on the screen. Preferably, the recitation aid is operable to align the text of the Quran which is displayed on the screen with the words of the Quran as they are recognised by the speech recognition module. Conveniently, the recitation aid is operable to align the text of the Quran displayed on the screen by turning a page of the Quran displayed on the screen.

Advantageously, the recitation aid is operable to align the text of the Quran displayed on the screen by scrolling the text of the Quran displayed on the screen.

According to another aspect of the present invention, there is provided a Quran recitation aid comprising: an audio input operable to receive an audio signal of a voice reciting text from the Quran, a speech recognition module operable to recognise words in the received audio signal, and a processing arrangement operable to compare the recognised words with text of the Quran and to provide an alert signal to a user if the recognised words do not follow the text of the Quran.

Preferably, the processing arrangement comprises an alignment algorithm which is operable to align the words recognised by the speech recognition module with text from the Quran. Conveniently, the alignment algorithm performs viterbi alignment.

Advantageously, the processing arrangement is operable to identify from the recognised words the school in which the Quran is recited. Preferably, the speech recognition module comprises a language recognition system which is trained using training data based on at least one sample of a correct recitation of at least part of the Quran. Conveniently, the recitation aid is a mobile device and the speech recognition module and the processing arrangement are implemented in one or more executable applications stored on the mobile device.

Advantageously, at least one of the speech recognition module and the processing arrangement is remote from the audio input.

Preferably, at least one of the speech recognition module and the processing arrangement is implemented in a remote server which is operable to communicate with a device incorporating the audio input.

According to the further aspect of the present invention, there is provided a method for aiding Quran recitation, the method comprising: displaying text from the Quran on a screen, receiving an audio signal of a voice reciting text from the Quran, recognising the words in the received audio signal using a speech recognition module, processing the recognised words to identify the position of the recognised words in the text of the Quran, and providing an indication of the recognised words in the text from the Quran which is displayed on the screen. Preferably, the step of providing the indication comprises highlighting the recognised words in the text from the Quran which is displayed on the screen.

Conveniently, the method further comprises providing an alert signal if the recognised words do not appear in the text of the Quran. Advantageously, the method further comprises providing an alert signal if the order of the words recognised by the speech recognition module do not follow the order in which the words appear in the text of the Quran. Preferably, the method further comprises highlighting the words of the Quran that should have been spoken differently in the text of the Quran which is displayed on the screen.

Conveniently, the method further comprises storing a data file representing the history of the accuracy of the Quran recitation.

Advantageously, the method further comprises displaying the history of the accuracy of a recitation by highlighting the correctly recited words in the text of the Quran on the screen and highlighting the incorrectly recited words differently in the text of the Quran on the screen.

Preferably, the method further comprises aligning the text of the Quran which is displayed on the screen with the words of the Quran as they are recognised by the speech recognition module.

Conveniently, the method comprises aligning the text of the Quran displayed on the screen by turning a page of the Quran displayed on the screen.

Advantageously, the method comprises aligning the text of the Quran displayed on the screen by scrolling the text of the Quran displayed on the screen.

According to a still further aspect of the present invention, there is provided a method for aiding Quran recitation, the method comprising: receiving an audio signal of a voice reciting text from the Quran, recognising the words in the received audio signal using a speech recognition module, processing the recognised words to compare the recognised words with the text of the Quran, and providing provide an alert signal to a user if the recognised words do not follow the text of the Quran. Preferably, the step of processing the recognised words comprises aligning the recognised words with text from the Quran stored in a memory.

Conveniently, the method comprises computing a viterbi alignment. Advantageously, method further comprises identifying from the recognised words the school in which the Quran is recited.

According to a yet further embodiment of the invention, there is provided a system for aiding Quran recitation, the system comprising: an audio input operable to receive an audio signal of a voice reciting text from the Quran, a speech recognition module operable to recognise words in the received audio signal, a processing arrangement operable to identify the position of the recognised words in the text of the Quran, and a transmission module operable to transmit data representing the recognised words to a device which is configured to provide an indication of the recognised words to a user.

In order that the invention may be more readily understood, and so that further features thereof may be appreciated, embodiments of the invention will now be described, by way of example, with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram of a Quran recitation aid of an embodiment of the invention, Figure 2 is a schematic diagram of the internal components of the recitation aid shown in figure 1 , Figure 3 is a view corresponding to figure 1 but with incorrectly recited words highlighted on the screen, and Figure 4 is a flow diagram showing the operation of an embodiment of the invention.

Referring initially to figure 1 of the accompanying drawings, a Quran recitation aid in the form of a device 1 of an embodiment of the invention incorporates a screen 2 for displaying text from the Quran. Words from the Quran are represented schematically as dashes 3. Dashed lines 4 represent lines of text from the Quran.

The device 1 incorporates an audio input which, in this embodiment, is a microphone 5. In other embodiments the audio input is another type of audio input which may, for instance, be a line input to receive an analogue or digital signal representing a sound capture or recording.

In a preferred embodiment of the invention, the device 1 is a portable smart device which incorporates a battery, processor, memory and display screen. The memory stores computer readable instructions and an electronic version of the Quran.

Referring now to figure 2 of the accompanying drawings, components of the internal working of the device 1 are shown schematically. The device 1 incorporates a speech recognition module 6 which is connected to the microphone 5 to receive an audio signal provided by the audio input by the microphone 5. The speech recognition module comprises an alignment algorithm which is implemented in hardware or software for processing audio signals captured by the microphone 5 to recognise words in the audio signals. There are numerous speech recognition algorithms which are familiar to those skilled in the art that may be used with embodiments of the invention. The algorithm may be based on a viterbi algorithm derived using the Hidden Markov Model Toolkit (HTK) originating from Cambridge University. In one embodiment, the speech recognition algorithm incorporates an implementation of a viterbi alignment method which is preferably forced viterbi alignment.

In embodiments of the invention, the speech recognition module is trained using training data which is based on at least one known correct Quran recitation. The training data is preferably based on multiple Quran recitations so as to maximise the accuracy of the word recognition.

One embodiment of the invention is operable to identify the school, manner or method of recitation of the Quran. The speech recognition module is operable to recognise words in the Quran as it is recited in all established schools, manners or methods.

The speech recognition module recognises words in the audio signal received by the microphone 5 and provides data representing a transcription of the recognised words to a processing arrangement in the form of processor 7. The processor 7 is operable to identify the position of the recognised words in the text of the Quran stored in a memory 8.

The processor identifies the recognised words by searching the text of the Quran for a word or string of words which matches one of the recognised words or a string of the recognised words. When the processor 7 identifies a match between the recognised word or words in the text of the Quran, the processor records the position of the identified text and provides an output to the screen 2 to indicate the recognised words on the screen 2. In a preferred embodiment of the invention, the screen 2 displays a page or a portion of the Quran text and indicates the recognised words by highlighting the words. In one embodiment, the highlighting is in the form of coloured text or a coloured background as represented by the hatched highlighted section 9 shown in figure 1 . In other embodiments, the recognised words are indicated on the screen 2 by changing the formatting of the words by, for instance, by showing the words in bold or underlined. In further embodiments, the recited words may be not highlighted on the screen at all, but the system still ensures that the recited words can be found in the displayed text on the screen.

The device 1 is configured to recognise words of the Quran as it is recited by the user or an Imam with the voice of the user or Imam being captured by the microphone 5. The device recognises the words and provides an indication of the recognised words on the screen 2 in substantial real-time as the words are spoken. The device 1 identifies to a user the words of the Quran that are being spoken and provides a near immediate indication of which line and verse of the Quran is being spoken. A Muslim can therefore bring the device 1 into group prayers which are already in progress and identify in real-time the written words of the Quran on the screen 2 as they are being spoken by the Imam.

The device 1 aligns the text of the Quran displayed on the screen with the text as it is spoken in real-time. The indication of the words on the screen 2 follows the words as they are spoken. The device 1 automatically flips the page of the Quran which is displayed on the screen 2 as the speaker moves on to the next page. The device 1 alternatively scrolls the text on the screen 2 down if the device 1 is set to operate using text scrolling instead of page flipping. The device 1 then continues to indicate the words on the subsequent page as they are spoken. In a further embodiment, the device 1 is operable to scroll the text of the Quran on the screen 2 so that the current words of the Quran which are being spoken are displayed on the screen 2. The automatic flipping of pages or scrolling of text on the screen 2 allows the device 1 to display the words of the Quran as they are spoken in real-time without the user having to interact with the device 1 . The user can simply look at the screen 2 to follow the written Quran as it is spoken.

The device 1 is operable to monitor the alignment between the recognised words and the spoken words in the text of the Quran. The device 1 is operable to look ahead in the stored Quran text to determine the next words of the Quran which are due to be spoken. If the device 1 detects that spoken words of the Quran which are received by the microphone do not follow on from previous words of the Quran then the device 1 detects the misalignment and provides an indication to the user that there has been a mistake in the recitation.

In a preferred embodiment, the device 1 indicates the mistake by highlighting the incorrectly recited words on the screen 2 in a different format to the highlighting used for the correctly recited words. For instance, the device 1 may highlight correctly recited words in green and incorrectly recited words in red. An example of the different indication for incorrectly recited words is shown in figure 3 by the hatched region 10.

A user is able to identify the mistake in the recitation quickly by noticing the indication of the incorrectly recited words on the screen 2. The user can then speak out loud the words which should have been recited in order to prompt the Imam to correct his recitation.

In a further embodiment, the device 1 is operable to indicate incorrectly recited words by outputting an alert sound. In one embodiment, the alert sound is a recording of a correct recitation of the words of the Quran. A still further embodiment does not incorporate a screen for displaying the Quran text. This embodiment provides an alert if the device detects mistakes in a recitation.

Another embodiment is implemented in a smartphone or another smart device which is configured to alert a user to recitation mistakes. The user does not have to carry the smart device in his hand. The user can simply put the smart device in his pocket or close to him to hear an alert produced by the smart device when there is a mistake in a recitation. Alternatively, the device can connect with an earpiece to be worn by a user to play back the aya or verse to the user if the Imam makes a mistake in his recitation.

In one embodiment, the device 1 is operable to retain a data file as a record of the history of the indication of the correctly and incorrectly recited words for subsequent review by a user. A user can review the stored data on the screen 2 to see which words were recited incorrectly. A user can thus use the device 1 as a learning tool when the user recites the Quran from memory so that the user knows where he or she has made a mistake. The user can then go back and memorise the passage that they got wrong. One embodiment of the invention can be used by any Muslim to monitor his/her accuracy in reciting the Quran to in improving his/her memorisation of the Quran.

The operation of an embodiment of the invention will now be described with reference to figure 4. A spoken recitation of the Quran is captured by the device 1 via the microphone 5. At an in itial isation step 1 1 , the speech recognition module 6 is activated along with the processor 7 to align the recognised words of the recited Quran with the Quran stored in the memory 8. The system checks the alignment and attempts to realign the recited words with the stored words until the two are aligned. At step 1 2, the system tracks the alignment of the recited words with the spoken words of the Quran stored in the memory 8. The system monitors the alignment and if the two are aligned then the system checks whether the end of the page has been reached at step 13. If the end of the page has not been reached then the system assesses the confidence of the recognition of the spoken words at step 14. The system highlights the words on the screen 2 in a particular colour or formatting at step 15 to indicate if the spoken words are correct or incorrect.

If, at step 13, the system determines that the recitation has reached the end of a page then the system flips the page at step 16 and displays the subsequent page on the screen 2.

Embodiments of the invention are described above as being implemented in a mobile smart device. However, it is to be appreciated that some of the features of the embodiments may be implemented in a separate device which is remote from the smart device 1 . For instance, in one embodiment, at least one of the speech recognition module 6, processor 7 and memory 8 are im plemented in a remote device, such as a remote server. In this embodiment, the device 1 is equipped with a communication arrangement to communicate with the remote server.

The device 1 is operable to capture sound and transmit data representing the sound such as an audio feature vector to the remote server where it is processed by the speech recognition module. The remote server then sends instruction data to the device 1 to display the words of the Quran on the screen 2 with the highlighting as described above. In this further embodiment, the remote server carries out some or all of the processing in order to reduce the processing load on the device 1 . In a yet further embodiment, the remote server is operable to transmit data to a plurality of devices to enable multiple users following the same Quran recitation to follow the words of the Quran on the devices as the words are spoken. In a still further embodiment of the invention, the speech recognition module and other elements of the embod iment are implemented as a computer program product. In one embodiment, the computer program product is provided in the form of an application or "app" which may be installed on a mobile smart device, such as a smartphone or tablet. The application may alternatively be installed on a desktop or laptop computer or dedicated hardware designed to run the application.

When used in this specification and claims, the terms "comprises" and "comprising" and variations thereof mean that the specified features, steps or integers are included. The terms are not to be interpreted to exclude the presence of other features, steps or components.

Claims

CLAIMS:

1 . A Quran recitation aid comprising:

a screen for displaying text from the Quran,

an audio input operable to receive an audio signal of a voice reciting text from the Quran,

a speech recognition module operable to recognise words in the received audio signal, and

a processing arrangement operable to identify the position of the recognised words in the text of the Quran and to provide an indication of the recognised words on the screen.

2. A Quran recitation aid according to claim 1 , wherein the aid is operable to indicate the recognised words on the screen by highlighting the words in the text from the Quran which is displayed on the screen.

3. A Quran recitation aid according to claim 1 or claim 2, wherein the processing arrangement is operable to provide an alert signal if the recognised words do not appear in the text of the Quran.

4. A Quran recitation aid according to any one of the preceding claims, wherein the processing arrangement is operable to provide an alert signal if the order of the words recognised by the speech recognition module do not follow the order in which the words appear in the text of the Quran.

5. A Quran recitation aid according to claim 4 as dependent on claim 2, wherein the processing arrangement provides the alert signal by highlighting the words of the Quran that should have been spoken differently in the text of the Quran which is displayed on the screen.

6. A Quran recitation aid according to claim 4 or claim 5, wherein the recitation aid is operable to store a data file representing the history of the accuracy of the Quran recitation.

7. A Quran recitation aid according to claim 6, wherein the recitation aid is operable to display the history of the accuracy of a recitation by highlighting the correctly recited words in the text of the Quran on the screen and highlighting the incorrectly recited words differently in the text of the Quran on the screen.

8. A Quran recitation aid according to any one of the preceding claims, wherein the recitation aid is operable to align the text of the Quran which is displayed on the screen with the words of the Quran as they are recognised by the speech recognition module.

9. A Quran recitation aid according to claim 8, wherein the recitation aid is operable to align the text of the Quran displayed on the screen by turning a page of the Quran displayed on the screen.

10. A Quran recitation aid according to claim 9, wherein the recitation aid is operable to align the text of the Quran displayed on the screen by scrolling the text of the Quran displayed on the screen.

1 1 . A Quran recitation aid comprising:

a processing arrangement operable to compare the recognised words with text of the Quran and to provide an alert signal to a user if the recognised words do not follow the text of the Quran.

12. A Quran recitation aid according to any one of the preceding claims, wherein the processing arrangement comprises an alignment algorithm which is operable to align the words recognised by the speech recognition module with text from the Quran.

13. A Quran recitation aid according to claim 12, wherein the alignment algorithm performs viterbi alignment.

14. A Quran recitation aid according to any one of the preceding claims, wherein the processing arrangement is operable to identify from the recognised words the school in which the Quran is recited.

15. A Quran recitation aid according to any one of the preceding claims, wherein the speech recognition module comprises a language recognition system which is trained using training data based on at least one sample of a correct recitation of at least part of the Quran.

16. A Quran recitation aid according to any one of the preceding claims, wherein the recitation aid is a mobile device and the speech recognition module and the processing arrangement are implemented in one or more executable applications stored on the mobile device.

17. A Quran recitation aid according to any one of claims 1 to 15, wherein at least one of the speech recognition module and the processing arrangement is remote from the audio input.

18. A Quran recitation aid according to claim 17, wherein at least one of the speech recognition module and the processing arrangement is implemented in a remote server which is operable to communicate with a device incorporating the audio input.

19. A method for aiding Quran recitation, the method comprising:

displaying text from the Quran on a screen,

receiving an audio signal of a voice reciting text from the Quran, recognising the words in the received audio signal using a speech recognition module,

processing the recognised words to identify the position of the recognised words in the text of the Quran, and

providing an indication of the recognised words in the text from the Quran which is displayed on the screen.

20. A method according to claim 19, wherein the step of providing the indication comprises highlighting the recognised words in the text from the Quran which is displayed on the screen.

21 . A method according to claim 19 or claim 20, wherein the method further comprises providing an alert signal if the recognised words do not appear in the text of the Quran.

22. A method according to any one of claims 19 to 21 , wherein the method further comprises providing an alert signal if the order of the words recognised by the speech recognition module do not follow the order in which the words appear in the text of the Quran.

23. A method according to claim 22 as dependent on claim 20, wherein the method further comprises highlighting the words of the Quran that should have been spoken differently in the text of the Quran which is displayed on the screen.

24. A method according to any one of claims 19 to 23, wherein the method further comprises storing a data file representing the history of the accuracy of the Quran recitation.

25. A method according to claim 24, wherein the method further comprises displaying the history of the accuracy of a recitation by highlighting the correctly recited words in the text of the Quran on the screen and highlighting the incorrectly recited words differently in the text of the Quran on the screen.

26. A method according to any one of claims 19 to 25, wherein the method further comprises aligning the text of the Quran which is displayed on the screen with the words of the Quran as they are recognised by the speech recognition module.

27. A method according to claim 26, wherein the method comprises aligning the text of the Quran displayed on the screen by turning a page of the Quran displayed on the screen.

28. A method according to claim 26, wherein the method comprises aligning the text of the Quran displayed on the screen by scrolling the text of the Quran displayed on the screen.

29. A method for aiding Quran recitation, the method comprising:

processing the recognised words to compare the recognised words with the text of the Quran, and

providing provide an alert signal to a user if the recognised words do not follow the text of the Quran.

30. A method according to any one of claims 19 to 29, wherein the step of processing the recognised words comprises aligning the recognised words with text from the Quran stored in a memory.

31 . A method according to claim 30, wherein the method comprises computing a viterbi alignment.

32. A method according to any one of claims 19 to 31 , wherein method further comprises identifying from the recognised words the school in which the Quran is recited.

33. A system for aiding Quran recitation, the system comprising:

a speech recognition module operable to recognise words in the received audio signal,

a processing arrangement operable to identify the position of the recognised words in the text of the Quran, and

a transmission module operable to transmit data representing the recognised words to a device which is configured to provide an indication of the recognised words to a user.