Alphanumeric message composing method using telephone keypad

• US 6,052,443 A
• Filed: 10/05/1998
• Issued: 04/18/2000
• Est. Priority Date: 05/14/1998
• Status: Expired due to Term

• Current Assignee: Google Technology Holdings LLC (Alphabet Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Perera, Surendra, Chang, Lu, Huang, Jian-Cheng, Schwendeman, Robert J.
• Primary Examiner(s): Chan, Wing F.
• Application Number:
  US09/166,739
• Time in Patent Office: 561 Days
• Field of Search: 379/93.01, 379/93.14, 379/93.17, 379/93.18, 379/93.22-93.28, 379/93.37, 379/88.15, 379/88.16, 379/88.24, 379/88.26, 455/31.2, 455/31.3, 455/38.1, 455/38.5, 340/311.1, 340/825.44
• US Class Current: 379/93.18
• CPC Class Codes: G06F 3/0237 using prediction or retrieval techniques
• CPC Class Codes: H04M 2250/70 methods for entering alphabetical characters, e.g. multi-tap or dictionary disambiguation methods for entering alphabetical characters per se G06F3/0237

Abstract

An interactive method for composing an alphanumeric message by a caller using a telephone keypad includes storing (215) a lexical database (135) from which unigram probabilities, forward conditional probabilities, and backward conditional probabilities for a plurality of words can be recovered; storing a received sequence of key codes (405) representing a sequence in which keys on a telephone style keypad are keyed; generating a word trellis including candidate words (415) derived from the sequence and the lexical database; determining a most likely phrase (420) from the candidate words, the unigram probabilities, forward conditional probabilities, and backward conditional probabilities; generating a most likely message (425) from the most likely phrase and presenting the most likely message to the caller; and confirming that the most likely message is the alphanumeric message (430).

Claims

1. An interactive method for composing an alphanumeric message using a telephone keypad, comprising in a messaging terminal the steps of:

storing a received sequence of key codes that identifies a sequence in which keys on a telephone style keypad are pressed by a caller in order to enter the alphanumeric message;

generating a word trellis from the sequence of key codes, the word trellis comprising at least one subset of candidate words recovered from a lexical database, each of the at least one subset of candidate words corresponding to one subsequence of the sequence of key codes;

determining a most likely phrase having a maximum combined probability determined from probabilities stored in the lexical database, that comprises candidate words that are a combination of one candidate word from each of the at least one subset of candidate words;

generating a most likely message from the most likely phrase and presenting the most likely message to the caller;

confirming that the most likely message is the alphanumeric message; and

editing the most likely message when, in response to the step of confirming, a message non-confirmation is received from the caller, wherein the step of editing the alphanumeric message further comprises the steps of;

prompting the caller to speak an edit command;

recognizing the edit command as one of a predetermined set of edit commands; and

performing the edit command using voice recognition that is governed by a subsequence of the sequence of key codes.

2. The method according to claim 1, wherein the step of recognizing the edit command comprises the step of recognizing a word edit command, and wherein the step of performing the edit command further comprises the steps of:

prompting for a vocalized word in response to the word edit command;

recording the vocalized word;

performing voice recognition using a small word recognizer that matches the vocalized word to a candidate word of the at least one subset of candidate words corresponding to the subsequence.

3. The method according to claim 2, wherein the step of performing the edit command further comprises the steps of:

prompting for a sequence of vocalized letters in response to a failure to match the vocalized word to any candidate word of the at least one subset of candidate words corresponding to the subsequence;

recording the sequence of vocalized letters; and

performing voice recognition using a alphabet recognizer that matches the sequence of vocalized letters to the subsequence of the sequence of keycodes.

4. The method according to claim 1, wherein the step of recognizing an edit command comprises the step of recognizing a spelling command for a word of the alphanumeric message, and wherein the step of performing the edit command further comprises the steps of:

prompting for a sequence of vocalized letters in response to the spelling command;

recording the sequence of vocalized letters; and

performing voice recognition using a alphabet recognizer that matches the sequence of vocalized letters to the subsequence of the sequence of keycodes.

5. An interactive method for composing an alphanumeric message using a telephone keypad, comprising in a messaging terminal the steps of:

storing a received sequence of key codes that identifies a sequence in which keys on a telephone style keypad are pressed by a caller in order to enter the alphanumeric message;

generating a word trellis from the sequence of key codes, the word trellis comprising at least one subset of candidate words recovered from a lexical database;

determining a most likely phrase having a maximum combined probability determined from probabilities stored in the lexical database, that comprises candidate words that are a combination of one candidate word from each of the at least one subset of candidate words;

generating a most likely message from the most likely phrase and presenting the most likely message to the caller;

confirming that the most likely message is the alphanumeric message; and

editing the most likely message when, in response to the step of confirming, a message non-confirmation is received from the caller, the editing performed in response to entries made by an operator.

6. The method according to claim 5, wherein the step of editing the most likely message comprises the steps of:

determining one or more incorrect candidate words of the most likely message and determining all remaining candidate words as confirmed candidate words of the most likely message when, in response to the step of confirming, a message non-confirmation is received from the caller;

presenting a request to the caller to vocalize spoken words of the alphanumeric message that correspond to each of one or more incorrect candidate words, in response to the step of determining one or more incorrect candidate words;

recording the spoken words; and

presenting to the operator the confirmed candidate words and the spoken words.

7. The method according to claim 6, wherein the step of editing the alphanumeric message further comprises the step of:

generating an operator edited version of the alphanumeric message in response to entries made by the operator that are generated after the step of presenting to the operator candidate words and the spoken words.

8. The method according to claim 7, wherein the operator entries are keyboard entries.

9. The method according to claim 6, wherein the step of editing the alphanumeric message further comprises the steps of:

establishing a voice link between the caller and the operator after the step of presenting to the operator the confirmed candidate words and the spoken words; and

generating an operator edited version of the alphanumeric message in response to entries made by the operator after the step of establishing a voice link.

10. The method according to claim 9, wherein the voice link is automatically established by the messaging terminal after the step of presenting to the operator the confirmed candidate words and the spoken words.

11. The method according to claim 9, wherein the voice link is established by a command generated by the operator.