Method for compressing dictionary data

• US 7,181,388 B2
• Filed: 11/11/2002
• Issued: 02/20/2007
• Est. Priority Date: 11/12/2001
• Status: Expired due to Fees

• Current Assignee: Nokia Corporation
• Original Assignee: Nokia Corporation
• Inventors: Tian, Jilei
• Primary Examiner(s): {hacek over (S)}mits; Talivaldis Ivars
• Assistant Examiner(s): Ng; Eunice
• Application Number:
  US10/292,122
• Publication Number:
  US 20030120482A1
• Time in Patent Office: 1,562 Days
• Field of Search: 704/1, 704/10, 704/256
• US Class Current: 704/10
• CPC Class Codes: G10L 15/12 using dynamic programming techniques, e.g. dynamic time warping [DTW]
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

The invention relates to pre-processing of a pronunciation dictionary for compression in a data processing device, the pronunciation dictionary comprising at least one entry, the entry comprising a sequence of character units and a sequence of phoneme units. According to one aspect of the invention the sequence of character units and the sequence of phoneme units are aligned using a statistical algorithm. The aligned sequence of character units and aligned sequence of phoneme units are interleaved by inserting each phoneme unit at a predetermined location relative to the corresponding character unit.

Claims

1. A method for pre-processing a pronunciation dictionary for compression in a data processing device, the pronunciation dictionary comprising at least one entry, the entry comprising a sequence of character units and a sequence of phoneme units,

the method comprising:

aligning said sequence of character units and said sequence of phoneme units using a statistical algorithm so that the alignment between said character units and said phoneme units is determined; and

interleaving said aligned sequence of character units and said aligned sequence of phoneme units by inserting each phoneme unit at a predetermined location relative to the corresponding character unit.

2. The method of claim 1, wherein said alignment is determined by employing the statistical algorithm, a HMM-Viterbi algorithm.

3. The method of claim 1, wherein said phoneme units are located next to corresponding character units.

4. The method of claim 1, wherein said aligned sequence of character units and said aligned sequence of phoneme units are made to include an equal number of units by at least one of the following insertions:

inserting graphemic epsilons to said sequence of character units; and

inserting phonemic epsilons into said sequence of phoneme units.

5. The method of claim 1, wherein said character units are letters or white space characters.

6. The method of claim 1, wherein said phoneme units are letters or whitespace characters representing a single phoneme or a phonemic epsilon and one said unit is denoted by at least one character.

7. The method of claim 1, the method further comprising:

mapping each phoneme unit into one symbol.

8. A computer program product loadable into the memory of a data processing device, comprising a code which is executable in the data processing device causing the data processing device to:

retrieve from the memory a pronunciation dictionary comprising at least one entry, the entry comprising a sequence of character units and a sequence of phoneme units;

align said sequence of character units and said sequence of phoneme units using a statistical algorithm; and

interleave said aligned sequence of character units and said aligned sequence of phoneme units by inserting each phoneme unit at a predetermined location relative to the corresponding character unit.

9. A data processing device comprising memory for storing a pronunciation dictionary comprising at least one entry, the entry comprising a sequence of character units and a sequence of phoneme units, wherein

the data processing device is configured to retrieve from the memory a pronunciation dictionary comprising at least one entry;

the data processing device is configured to align said sequence of character units and said sequence of phoneme units using a statistical algorithm; and

the data processing device is configured to interleave said aligned sequence of character units and said aligned sequence of phoneme units by inserting each phoneme unit at a predetermined location relative to the corresponding character unit.

10. The data processing device of claim 9, wherein the data processing device is configured to determine said alignment by employing the statistical algorithm, a HMM-Viterbi algorithm.

11. The data processing device of claim 9, wherein the data processing device is configured to locate said phoneme units next to corresponding character units.

12. The data processing device of claim 9, wherein the data processing device is configured to cause said aligned sequence of character units and said aligned sequence of phoneme units include an equal number of units by at least one of the following insertions:

inserting graphemic epsilons to said sequence of character units; and

inserting phonemic epsilons into said sequence of phoneme units.

13. The data processing device of claim 9, wherein said character units are letters or whitespace characters.

14. The data processing device of claim 9, wherein said phoneme units are letters or whitespace characters representing a single phoneme or a phonemic epsilon and one said unit is denoted by at least one character.

15. The data processing device of claim 9, wherein the data processing device is configured to map each phoneme unit into one symbol.