Verifying a speaker using random combinations of speaker's previously-supplied syllable units

1. A method for verifying a speaker by a computer, comprising:

randomly selecting a code word by assembling previously stored syllable units, the syllable units having been previously supplied by the speaker;

allowing a speaker to utter a code word, the code word having at least two syllable units with at least one syllable unit being a component of another code word;

comparing the code word uttered by the speaker with a pattern formed of previously stored syllable units having been previously supplied by the speaker; and

verifying the speaker only if the pattern matches the uttered code word with a predetermined minimum quality.

2. The method as claimed in claim 1 , further comprising:

requesting the speaker to utter the code word, and

displaying to the speaker the code word to be uttered.

3. The method as claimed in claim 1 , wherein the speaker is only authenticated when he has uttered a number of code words in each case with sufficient minimum quality.

4. The method as claimed in claim 1 , wherein each syllable unit comprises at least one number.

5. The method as claimed in claim 1 , wherein the pattern comprises two-digit numbers.

6. The method as claimed in claim 1 , wherein the pattern comprises two-digit numbers within a range from 21 to 99.

7. The method as claimed in claim 2 , wherein the speaker is only authenticated when he has uttered a number of code words in each case with sufficient minimum quality.

8. The method as claimed in claim 7 , wherein each syllable unit comprises at least one number.

9. The method as claimed in claim 8 , wherein the pattern comprises two-digit numbers.

10. The method as claimed in claim 9 , wherein the pattern comprises two-digit numbers within a range from 21 to 99.

11. A method for verifying a speaker, comprising:

assembling a new code word from syllable units;

allowing the speaker to utter the code word;

comparing the code word uttered by the speaker with at least one pattern, formed of previously stored syllable units, the syllable units having been previously supplied by the speaker; and

verifying the speaker only if the at least one pattern matches the uttered code word with a predetermined minimum quality

wherein the code word is randomly selected by assembling the previously stored syllable units having been previously supplied by the speaker.

12. The method as claimed in claim 11 , further comprising:

requesting the speaker to utter the code word, and

displaying to the speaker the code word to be uttered to the speaker.

13. The method as claimed in claim 11 , wherein the speaker is only authenticated when he has uttered a number of code words in each case with sufficient minimum quality.

14. The method as claimed in claim 11 , wherein each syllable unit comprises at least one number.

15. The method as claimed in claim 11 , wherein the pattern comprises two-digit numbers.

16. The method as claimed in claim 11 , wherein the pattern comprises two-digit numbers within a range from 21 to 99.

17. The method as claimed in claim 12 , wherein the speaker is only authenticated when he has uttered a number of code words in each case with sufficient minimum quality.

18. The method as claimed in claim 17 , wherein each syllable unit comprises at least one number.

19. The method as claimed in claim 18 , wherein the pattern comprises two-digit numbers.

20. The method as claimed in claim 19 , wherein the pattern comprises two-digit numbers within a range from 21 to 99.

21. A method for training patterns for verifying a speaker, comprising:

allowing the speaker to speak a plurality of words;

automatically segmenting the words into syllable units;

storing the syllable units;

randomly selecting a code word by assembling the syllable units having been previously supplied by the speaker to form the code word; and

using the code word, including the syllable units having been previously supplied by the speaker, to verify the speaker.

22. The method as claimed in claim 21 , wherein the words used to verify the speaker is a pattern assembled from the segmented syllable units.

23. The method as claimed in claim 21 , wherein the words are automatically segmented into syllable units by a DTW algorithm.

24. The method as claimed in claim 21 , wherein the words are automatically segmented using a plurality of predetermined voice samples.

25. The method as claimed in claim 21 , wherein voice samples for a particular speaker replace predetermined generic voice samples.

26. The method as claimed in claim 22 , wherein the words are automatically segmented into syllable units by a DTW algorithm.

27. The method as claimed in claim 26 , wherein the words are automatically segmented using a plurality of predetermined voice samples.

28. The method as claimed in claim 27 , wherein voice samples for a particular speaker replace predetermined generic voice samples.

29. A computer readable medium storing a program for controlling a computer to perform a method comprising:

allowing a speaker to utter at least one code word, the code word having at least two syllable units with at least one syllable unit being a component of another code word;

comparing the code word uttered by the speaker with at least one pattern formed of previously stored syllable units, the syllable units having been previously supplied by the speaker; and

verifying the speaker only if the at least one pattern matches the uttered code word with a predetermined minimum quality,

wherein the code word is randomly selected by assembling the previously stored syllable units having been previously supplied by the speaker.