Name and address processor

• US 5,761,640 A
• Filed: 12/18/1995
• Issued: 06/02/1998
• Est. Priority Date: 12/18/1995
• Status: Expired due to Term

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: NYNEX Science & Technology, Inc. (Verizon Communications Inc.)
• Inventors: Kalyanswamy, Ashok, Man, Edward
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US08/574,233
• Time in Patent Office: 897 Days
• Field of Search: 395/2.69, 395/2.86, 395/2.79, 395/2.09, 395/2.75, 395/795, 395/759, 379/142, 379/127, 704/4, 704/9, 704/258
• US Class Current: 704/260
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination

Abstract

A name and address processor for processing text contained within an existing database for subsequent text-to-speech synthesis. The processor receives as input a listing contained within a textual source database, intelligently recognizes any fields contained within the textual source, normalizes the text contained within the fields, detects acronyms contained within the fields, identifies and marks any particular textual entries as necessitating spelling and then formats the processed text for output to a text-to-speech synthesizer. The processor processes in parallel all name field entries, address field entries, and locality field entries using tables of rules as well as both regular expression and non-regular expression methodologies.

Claims

1. A method for processing text contained within a database for subsequent synthesis by a text-to-speech synthesizer comprising the steps of:

inputting a listing from a database containing the text to be processed;

parsing the text into one or more distinct fields;

processing in parallel and generating an output for each of the distinct fields wherein said parallel processing includes the steps of;

i) normalizing the text contained within each of the fields utilizing both regular expressions to normalize the text and non-regular expressions to normalize the text;

ii) detecting acronyms contained within the text;

iii) identifying text which is to be spelled-out by the text-to-speech synthesizer; and

combining the output of each of the parallel processing steps into a single output, for presentation to the text-to-speech synthesizer.

2. The method according to claim 1 wherein said parsing step produces a Name Field, an Address Field and a Locality Field.

3. The method according to claim 1 wherein said step of normalizing the text contained in each of the fields includes a sub-step of checking for embedded numbers.

4. A device for processing textual data contained within a database for subsequent synthesis by a text-to-speech synthesizer such that resultant speech is enhanced, said device comprising:

a computer processor;

a control module including at least one application for execution by the computer processor;

a collection of processing tables and processing rules for use by the computer processor in processing the textual data within the database;

a start up module in communication with said control module and said collection of tables and rules, for execution by the computer processor to initialize said tables prior to processing said text;

a configuration file for execution by the computer processor to configure the at least one application;

a set of tools in communication with said at least one application, said tables and rules and said configuration file, said set of tools including;

an intelligent field recognizer for generating a plurality of fields of text from the textual data contained within the database;

a plurality of field normalizer modules, one for each field generated, for normalizing the fields of text generated by the intelligent field recognizer;

an acronym detector module for detecting acronyms contained within the normalized fields of text generated by the plurality of field normalizer modules;

means, in communication with the at least, one application and the tables a rules, for determining whether the textual data is a business listing or a residence listing; and

an output formatter for generating formatted fields of text after the fields of text have been normalized by the field normalizers and have had acronyms detected by the acronym detector;

wherein said formatted fields of text are presented to the text-to-speech synthesizer for producing speech corresponding to the textual data processed.

5. The device according to claim 4 wherein said plurality of normalizer modules further comprise a Name Field text normalizer module, an Address Field text normalizer module and a Locality Field text normalizer module.

6. The device according to claim 5 wherein said Name Field text normalizer module uses a data structure which comprises:

phone_-- num_-- info, telephone_-- num, synthesizer_-- name, listing_-- name, family_-- name, given_-- name, DBA_-- link, care_-- of_-- link, attention_-- link, directive_-- text and listing_-- type.

7. The device according to claim 5 wherein said Address Field text normalizer module uses a data structure which comprises:

a telephone_-- num, address, house_-- num, streetname, street_-- type and street_-- suffix.

8. The device according to claim 5 wherein said Locality Field text normalizer module uses a data structure which comprises:

telephone_-- num, city, state, zip_-- code, and zip_-- plus_-- four.