System for storing voice recognizable identifiers using a limited input device such as a telephone key pad

• US 6,728,348 B2
• Filed: 11/30/2000
• Issued: 04/27/2004
• Est. Priority Date: 11/30/2000
• Status: Expired due to Fees

• Current Assignee: Comverse Incorporated (Mavenir Group Holdings Ltd.)
• Original Assignee: Comverse Incorporated (Mavenir Group Holdings Ltd.)
• Inventors: Schmandt, Christopher M., Denenberg, Lawrence A.
• Primary Examiner(s): Tsang, Fan
• Assistant Examiner(s): Sing, Simon P.
• Application Number:
  US09/725,676
• Publication Number:
  US 20020064257A1
• Time in Patent Office: 1,244 Days
• Field of Search: 379/88.02, 379/88.03, 379/88.04, 379/218.01, 379/93.18, 704/251
• US Class Current: 379/93.18
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: H04M 1/271 controlled by voice recognition

Abstract

The present invention is a system that uses a limited text input entry device to narrow the possibilities for the selection of a phonetically based name used in a voice dialer. The user enters a DTMF signature of a name using only one key for each letter of the name. The signature is used by a dictionary to generate likely possibilities for the word. The user is asked whether the name is the name entered via playing the name to the user. If the user confirms the choice, the phonetic representation is stored along with a corresponding telephone number. If the played word does not match, a word generator generates additional likely possibilities from the signature. A speech recognizer picks the best representation from the dictionary and generator words based on a spoken version of the name. A first name and last name procedure, in which the first and last name signatures produce separate candidate lists, is used to provide an entry to an address book for voice dialing.

Claims

1. A method, comprising:

producing phonetic word candidates using a multiple word possibility input sequence of a word, according to a word dictionary and a score-based word generator that generates phonetically possible words based upon the input sequence; and

selecting, using a speech recognizer, one of the produced word candidates for storage based upon the produced phonetic word candidates and using a spoken version of the input word during the selecting.

2. A method as recited in claim 1, wherein the multiple word possibility input sequence comprises a DTMF sequence.

3. A method as recited in claim 1, wherein the multiple word possibility input sequence comprises a verbally spelled sequence.

4. A method as recited in claim 1, wherein the input sequence is a multiple name possibility DTMF sequence.

5. A method as recited in claim 4, wherein the producing comprises:

producing a first set of name candidates using a dictionary; and

generating a second set of name candidates using a score based name generator.

6. A method as recited in claim 5, wherein the producing further comprises:

presenting a highest probability dictionary candidate to a user for selection; and

storing the highest probability candidate as the phonetic based spelling of the multiple name possibility DTMF sequence when selected by the user.

7. A method as recited in claim 6, further comprising:

storing a telephone number corresponding to the phonetic based spelling; and

performing voice dialing operations using the phonetic based spelling and the telephone number.

8. A method as recited in claim 4, wherein the producing comprises producing a set of name candidates using a dictionary.

9. A method as recited in claim 4, wherein the producing comprises generating a set of name candidates using a score based name generator.

10. A method as recited in claim 1, wherein input sequence comprises first and last name signatures.

11. A method as recited in claim 10, wherein the producing comprises:

producing first and last name dictionary candidates using first and last name dictionaries; and

producing first and last name generated candidates using first and last name score based name generators.

12. A method as recited in claim 11, wherein the producing further comprises:

presenting a highest probability first and last name dictionary candidates to a user for selection; and

storing the highest probability candidates as the corresponding phonetic based spellings of the multiple name possibility DTMF sequences as selected by the user.

13. A method of voice dialing, comprising:

inputting by a user a multiple name possibility single key DTMF input sequence comprising first and last name signatures via a DTMF telephone key pad;

producing phonetic name candidates using the input sequence, said producing comprising;

producing a first set of first and last name candidates using first and last name dictionaries responsive to the signatures; and

generating a second set of first and last name generated candidates using first and last name score based name generators responsive to the signatures;

selecting one of the produced candidates from the produced first and second sets of name candidates for storage using a spoken version of the name by the user during the selecting using a speech recognizer;

storing a telephone number corresponding to the stored candidate; and

performing recognition voice dialing using the stored candidate and the telephone number.

14. The method of claim 13, wherein the producing further comprises:

before the selecting audibly presenting a highest probability first and last name dictionary candidate from the first set to the user for selection; and

storing the highest probability candidate as the selected phonetic based spelling of the multiple name possibility DTMF sequence, if selected by the user.

15. An apparatus, comprising:

a DTMF telephone inputting a multiple name possibility DTMF sequence as a name and a spoken version of the sequence spoken by a user; and

a computer producing phonetic name dictionary candidates using a multiple word possibility input sequence as input to both a phonetic dictionary and a score based name generator, and selecting one of the produced name candidates for storage using the user spoken version of the sequence.

16. The apparatus of claim 15, wherein the computer further allows the user to confirm whether a highest probability phonetic name from the dictionary candidates and from the generated name candidates using the score based name generator is the name, and performs the selecting if the highest probability name is not the name.

17. A computer readable storage storing at least one program and controlling a computer according to a process comprising:

producing phonetic word candidates using a multiple word possibility input sequence of a word, according to a word dictionary and a score-based word generator that generates phonetically possible words based upon the input sequence; and

selecting, using a speech recognizer, one of the word candidates for storage based upon the produced phonetic word candidates and using a spoken version of the input word by-The during the selecting.