Method and apparatus for conducting point-of-sale transactions using voice recognition

• US 5,758,322 A
• Filed: 12/09/1994
• Issued: 05/26/1998
• Est. Priority Date: 12/09/1994
• Status: Expired due to Fees

• Current Assignee: International Voice Register Incorporated
• Original Assignee: International Voice Register Incorporated
• Inventors: Rongley, Eric H.
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): Chawan, Vijay B.
• Application Number:
  US08/353,534
• Time in Patent Office: 1,264 Days
• Field of Search: 395/2.4, 395/2.6, 395/2.84, 395/2.52, 395/2.53, 395/2.55, 395/2.61, 395/2.56
• US Class Current: 704/275
• CPC Class Codes: G06Q 20/341 Active cards, i.e. cards including their own processing means, e.g. including an IC or chip
• CPC Class Codes: G06Q 20/4014 Identity check for transactions
• CPC Class Codes: G06Q 20/40145 Biometric identity checks
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition
• CPC Class Codes: G07F 7/1008 Active credit-cards provided with means to personalise their use, e.g. with PIN-introduction/comparison system
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A method and apparatus for conducting point-of-sale transactions using both speaker dependent and speaker independent voice recognition in which a spoken utterance from a first user is captured with a sound input device. The spoken utterance is compared to a first plurality of stored patterns to find a first match. A product order is generated based on the first match found in the comparing step. The comparing and generating steps may be repeated for each additional product purchased. A bill is then calculated based on the product orders. The plurality of stored patterns, also referred to as a voice layer, comprises a subset of a user vocabulary stored in a memory. The voice layer, i.e., the plurality of stored patterns to which the spoken utterances are compared, changes to include different sets of likely words from the vocabulary depending upon the context of the transaction. In the speaker dependent embodiment, the user vocabulary is specific to the first user.

Claims

1. A method for conducting a point-of-sale transaction using voice recognition, the transaction having a plurality of stages, the method comprising the steps of:

capturing a spoken utterance from a user with a sound input device;

comparing the spoken utterance to a plurality of stored patterns to find a match;

generating a first product order based on the match;

repeating the capturing, comparing, and generating steps for each additional spoken utterance; and

calculating a bill based on the product orders;

wherein the plurality of stored patterns comprises a subset of a vocabulary stored in a memory, the subset changing for selected comparing steps to include different predetermined sets of likely words from the vocabulary, each predetermined set of the likely words corresponding to one of the transaction stages, the vocabulary corresponding to the user;

changing the first product order to a second product in response to a first corrective utterance from the user;

incrementing an error count corresponding to a first stored pattern, the first stored pattern corresponding to the second product order;

prompting the user to repeat a second corrective utterance corresponding to the second product order for at least one repetition when the error count reaches a predetermined number;

capturing the at least one repetition of the second corrective utterance with the sound input device; and

combining the at least one repetition of the second corrective utterance with the first stored pattern to form a new pattern.

2. The method of claim 1 further comprising the steps of:

receiving a log-in command from the user; and

retrieving the vocabulary from a main memory in response to the vocal log-in command, the main memory having a plurality of user vocabularies stored therein, each vocabulary corresponding to a different user.

3. The method of claim 2 further comprising the step of determining access by the user to different levels of different operating modes based upon a security level associated with the user.

4. The method of claim 3 wherein the different operating modes are selected from a group consisting of order taking, inventory, timekeeping, goods receiving, and troubleshooting.

5. The method of claim 1 further comprising the steps of:

receiving a password from the user; and

clocking in the user in response to the password.

6. The method of claim 1 further comprising the steps of:

receiving a vocal clock-out command from the user via the sound input device; and

clocking out the user in response to the vocal clock-out command.

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

prompting the user to speak at least one additional utterance;

capturing the at least one additional utterance with the sound input device; and

adding the at least one additional utterance to the plurality of stored patterns in the vocabulary.

8. The method of claim 1 further comprising the step of displaying the first product order on a display means.

9. The method of claim 1 further comprising the step of printing a printout corresponding to the product orders.

10. The method of claim 1 further comprising the step of opening a cash drawer in response to an open command from the user.

11. The method of claim 1 further comprising the steps of:

capturing a corrective utterance from the user with the sound input device;

deleting the first product order in response to the corrective utterance;

capturing a second utterance from the user with the sound input device;

comparing the second utterance to the plurality of stored patterns to find a second match; and

generating a second product order to replace the first product order based on the second match.

12. The method of claim 1 wherein each match found with the comparing step has a confidence level associated therewith, the method further comprising the step of prompting the user to enter a password when the confidence level associated with a particular match differs from previous confidence levels associated with the particular match by more than a first amount.

13. The method of claim 1 further comprising the steps of:

receiving a log in command from a first user;

retrieving a first user vocabulary corresponding to the first user from a main memory in response to the log in command, the main memory having a plurality of user vocabularies stored therein, each user vocabulary corresponding to a different user;

storing a location marker for the users active voice file;

capturing a log out command;

sending the current copy of the voice file back to the main memory;

updating the marker for the location of the active vocabulary for the user;

prompting a new user to repeat a plurality of test utterances in succession;

capturing the plurality of test utterances from the new user for storage in the main memory as part of the new user vocabulary, each test utterance being repeated by the new user at least once; and

storing the plurality of test utterances in the main memory, the test utterances corresponding to the new user vocabulary.

14. The method of claim 1 wherein there are a plurality of vocabularies stored in the memory representing different voice models, the method further comprising the step of selecting the first vocabulary from the plurality of vocabularies based on an analysis of at least a first utterance by the user.

15. The method of claim 13 further comprising directing the system to create a new user;

assigning at least one level of access rights to the new user; and

adding vocabulary words associated with each level of access to a vocabulary for the new user.