Speaker verification in a health monitoring system

• US 9,424,845 B2
• Filed: 08/26/2014
• Issued: 08/23/2016
• Est. Priority Date: 12/29/2011
• Status: Active Grant

• Current Assignee: Robert Bosch GmbH
• Original Assignee: Robert Bosch GmbH
• Inventors: Weng, Fuliang, Hasan, Taufiq, Feng, Zhe
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US14/468,411
• Publication Number:
  US 20140365219A1
• Time in Patent Office: 728 Days
• Field of Search: 704/246
• US Class Current: 1/1
• CPC Class Codes: G10L 17/10 Multimodal systems, i.e. based on the integration of multiple recognition engines or fusion of expert systems
• CPC Class Codes: G10L 17/22 Interactive procedures; Man-machine interfaces
• CPC Class Codes: G10L 17/24 the user being prompted to utter a password or a predefined phrase
• CPC Class Codes: G16H 40/20 for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms

Abstract

A method for verifying that a person is registered to use a telemedical device includes identifying an unprompted trigger phrase in words spoken by a person and received by the telemedical device. The telemedical device prompts the person to state a name of a registered user and optionally prompts the person to state health tips for the person. The telemedical device verifies that the person is the registered user using utterance data generated from the unprompted trigger phrase, name of the registered user, and health tips.

Claims

1. A system comprising:

a memory;

an audio input device configured to generate audio data from utterances spoken by a person;

a user interface device; and

an audio data processor operatively coupled to the memory, the audio input device, and the user interface device the audio data processor being configured to;

receive first audio data generated by the audio input device in response to a first utterance spoken by the person;

identify first utterance data in the first audio data;

store the first utterance data in the memory;

generate a prompt with the user interface device for the person to speak a registration name in response to the first utterance data corresponding to a predetermined trigger utterance;

receive second audio data generated by the audio input device in response to a second utterance spoken by the person in response to the prompt;

generate second utterance data in the second audio data corresponding to the registration name;

store the second utterance data in the memory; and

verify that the person is a user registered in a registration database stored in the memory in association with the registration name with reference to the first utterance data and the second utterance data.

2. The system of claim 1, the audio data processor being further configured to:

identify a first confidence score corresponding to a likelihood that the first utterance data and the second utterance data correspond to a first predetermined voice model stored in the memory, the first predetermined voice model being associated with the person;

identify a second confidence score corresponding to another likelihood that the first utterance data and the second utterance data correspond to a second predetermined voice model stored in the memory, the second predetermined voice model not being associated with the person; and

verify that the person is the user registered in the registration database in response to the first confidence score being greater than the second confidence score by at least a predetermined threshold.

3. The system of claim 1, the audio data processor being further configured to:

generate another prompt with the user interface device for the person to speak a predetermined phrase;

receive third audio data generated by the audio input device in response to a third utterance spoken by the person corresponding to the predetermined phrase;

identify third utterance data in the third audio data; and

verify that the person is the user registered in the registration database stored in the memory in association with the registration name with reference to the first utterance data, the second utterance data, and the third utterance data.

4. The system of claim 3 wherein the audio data processor generates a prompt for the person to speak a health tip as the predetermined phrase.

5. A method of operating a system comprising:

generating with an audio input device first audio data corresponding to utterances spoken by a person;

identifying with an audio data processor operatively connected to the audio input device first utterance data in the first audio data;

storing with the audio data processor the identified first utterance data in a memory;

generating with a user interface device a prompt for the person to speak a registration name in response to the first utterance data corresponding to a predetermined trigger utterance;

generating with the audio input device second audio data corresponding to utterances spoken by the person in response to the prompt;

identifying with the audio data processor second utterance data in the second audio data corresponding to the registration name;

storing with the audio data processor the second utterance data in the memory; and

verifying with the audio data processor that the person is a user registered in a registration database stored in the memory in association with the registration name with reference to the first utterance data and the second utterance data.

6. The method of claim 5, the verification further comprising:

identifying with the audio data processor a first confidence score corresponding to a likelihood that the first utterance data and the second utterance data correspond to a first predetermined voice model stored in the memory, the first predetermined voice model being associated with the person;

identifying with the audio data processor a second confidence score corresponding to another likelihood that the first utterance data and the second utterance data correspond to a second predetermined voice model stored in the memory, the second predetermined voice model not being associated with the person; and

verifying with the audio data processor that the person is the user registered in the registration database in response to the first confidence score being greater than the second confidence score by at least a predetermined threshold.

7. The method of claim 5 further comprising:

generating with the user interface device another prompt with the user interface device for the person to speak a predetermined phrase;

generating with the audio input device third audio data in response to a third utterance spoken by the person corresponding to the predetermined phrase;

identifying with the audio data processor third utterance data in the third audio data; and

verifying with the audio data processor that the person is the user registered in the registration database stored in the memory in association with the registration name with reference to the first utterance data, the second utterance data, and the third utterance data.

8. The method of claim 7 wherein the audio data processor generates a prompt for the person to speak a health tip as the predetermined phrase.