METHOD AND APPARATUS FOR LANGUAGE INDEPENDENT VOICE INDEXING AND SEARCHING

• US 20080162125A1
• Filed: 12/28/2006
• Published: 07/03/2008
• Est. Priority Date: 12/28/2006
• Status: Abandoned Application

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Li, Feipeng, Ma, Changxue C.
• Application Number:
  US11/617,265
• Publication Number:
  US 20080162125A1
• Time in Patent Office: Days
• US Class Current: 704/222
• CPC Class Codes: G06F 16/632 Query formulation
• CPC Class Codes: G06F 16/685 using automatically derived transcript of audio data, e.g. lyrics speech recognition G10L15/00
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units

Abstract

A method and apparatus for language independent voice searching in a mobile communication device is disclosed. The method may include receiving a search query from a user of the mobile communication device, converting speech parts in the search query into linguistic representations which covers at least one languages, generating a search phoneme lattice based on the linguistic representations, extracting query features from the search phoneme lattice, generating query feature vectors based on the extracted features, performing a coarse search using the query feature vectors and the indexing feature vectors from the indexing database, performing a fine search using the results of the coarse search and the indexing phoneme lattices stored in the indexing database, and outputting the fine search results to a dialog manager.

Claims

1. A method for language independent voice indexing and searching in a mobile communication device, comprising:

receiving a search query from a user of the mobile communication device;

converting speech parts in the search query into linguistic representations;

generating a search phoneme lattice based on the linguistic representations;

extracting query features from the generated search phoneme lattice;

generating query feature vectors based on the extracted query features;

performing a coarse search using the generated query feature vectors and indexing feature vectors from an indexing database, wherein the indexing database stores indices of indexing feature vectors from indexing phoneme lattices of audio files stored on the mobile communication devices;

performing a fine search using the results of the coarse search and the indexing phoneme lattices stored in the indexing database; and

outputting the fine search results to a dialog manager.

2. The method of claim 1, wherein the linguistic representations are at least one of words, morphemes, syllables, and phonemes of at least one language.

3. The method of claim 1, wherein the search query concerns an audio file stored on the mobile communication device.

4. The method of claim 3, wherein the audio file is one of audio recordings, voice mails, recorded conversations, notes, messages, and annotations.

5. The method of claim 1, wherein the coarse search generates a plurality of candidate audio files based on the search query.

6. The method of claim 5, wherein the fine search generates the best candidate out of the coarse search results.

7. The method of claim 1, wherein the mobile communication device is one of a mobile telephone, cellular telephone, a wireless radio, a portable computer, a laptop, an MP3 player, satellite radio, satellite television, Digital Video Recorder (DVR), and television set-top box.

8. An apparatus for language independent voice searching in a mobile communication device, comprising:

an indexing database that stores indices of indexing feature vectors from indexing phoneme lattices of audio files stored on the mobile communication devices; and

a voice search engine that receives a search query from a user of the mobile communication device, converts speech parts in the search query into linguistic representations, generates a search phoneme lattice based on the linguistic representations, extracts query features from the generated search phoneme lattice, generates query feature vectors based on the extracted query features, performs a coarse search using the query feature vectors and indexing feature vectors from the indexing database, performs a fine search using the results of the coarse search and the indexing phoneme lattices stored in the indexing database, and outputs the fine search results to a dialog manager.

9. The apparatus of claim 8, wherein the linguistic representations are at least one of words, morphemes, syllables, and phonemes of at least one language.

10. The apparatus of claim 8, wherein the search query concerns an audio file stored on the mobile communication device.

11. The apparatus of claim 10, wherein the audio file is one of audio recordings, voice mails, recorded conversations, notes, messages, and annotations.

12. The apparatus of claim 8, wherein the coarse search performed by the voice search engine generates a plurality of candidate audio files based on the search query.

13. The apparatus of claim 12, wherein the fine search performed by the voice search engine generates the best candidate out of the coarse search results.

14. The apparatus of claim 8, wherein the mobile communication device is one of a mobile telephone, cellular telephone, a wireless radio, a portable computer, a laptop, an MP3 player, satellite radio, satellite television, Digital Video Recorder (DVR), and television set-top box.

15. An apparatus for language independent voice searching in a mobile communication device, comprising:

an indexing database that stores indices of indexing feature vectors from indexing phoneme lattices of audio files stored on the mobile communication device;

a search automatic speech recognizer that receives a search query from a user of the mobile communication device and converts speech parts in the search query into linguistic representations;

a search phoneme lattice generator that generates a search phoneme lattice based on the linguistic representations;

a search feature vector generator that extracts query features from the search phoneme lattice and generates query feature vectors based on the extracted query features;

a coarse search module that performs a coarse search using the query feature vectors and indexing feature vectors from the indexing database; and

a fine search module that performs a fine search using the results of the coarse search and the indexing phoneme lattices stored in the indexing database, and outputs the fine search results to a dialog manager.

16. The apparatus of claim 15, wherein the linguistic representations are at least one of words, morphemes, syllables, and phonemes of at least one language.

17. The apparatus of claim 15, wherein the search query concerns an audio file stored on the mobile communication device.

18. The apparatus of claim 17, wherein the audio file is one of audio recordings, voice mails, recorded conversations, notes, messages, and annotations.

19. The apparatus of claim 15, wherein the coarse search module generates a plurality of candidate audio files based on the search query, and the fine search module generates the best candidate out of the coarse search results.

20. The apparatus of claim 15, wherein the mobile communication device is one of a mobile telephone, cellular telephone, a wireless radio, a portable computer, a laptop, an MP3 player, satellite radio, satellite television, Digital Video Recorder (DVR), and television set-top box.