Text editor for speech input

• US 4,914,704 A
• Filed: 10/30/1984
• Issued: 04/03/1990
• Est. Priority Date: 10/30/1984
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Cole, Alan G., Riekert, Robert H.
• Primary Examiner(s): Harkcom, Gary V.
• Assistant Examiner(s): Merecki, John A.
• Application Number:
  US06/666,212
• Time in Patent Office: 1,981 Days
• Field of Search: 381/41-46, 381/48, 364/513.5, 364/419, 364/200 MS File, 364/900 MS File
• US Class Current: 704/235
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G10L 2015/086 Recognition of spelled words

Abstract

A text editor is connected to a speech recognizing unit for editing preferably spoken input text using a display speech. For each text word (including digits), and each punctuation mark that can be recognized and is contained in a dictionary, a token is stored for holding information on character count, capitalization, left and right concatenation of the respective item, and for providing fields for context conditions. For each segment or entity recognized spoken text, a respective character string and associated token is transferred to storage in the editor to allow automatic formatting and correct displaying or printing of the text, including spaces and capitalization where required. Tokens are updated during editing to reflect modifications such as in the beginning of a sentence or in concatenation. Switching to spelling mode is provided for entering single spelled characters in case where a word cannot be recognized or where spelling is desired.

Claims

1. A system for editing information from a speech recognizing unit that provides coded identifiers of recognized speech segments, characterized in that the editing system includes:

dictionary and lookup means having inputs comprising said coded identifiers of speech segments each of which represents a word or punctuation word recognized by said speech recognizing unit, wherein said dictionary and lookup means is responsive to each coded identifier of a speech segment, to furnish for a speech segment a respective character string and a token associated with each character string, each token comprising a character count and concatenation indicia, the character count representing the number of characters in the character string, the concatenation indicia representing whether the character string is concatenated to a preceding or following character string,

editing means responsive to character strings and tokens furnished by said dictionary and lookup means, for generating formatted data including correct spacing and capitalization, and

display means responsive to said formatted data for displaying the speech segments in standard text form. display means responsive to said formatted data for displaying the speech segments in standard text form.

2. An editing system according to claim 1, characterized in that

said editing means comprises:

storing and formatting means for storing

(a) character strings furnished by said dictionary storage means and

(b) the tokens of a previously spoken text being edited,

means for entering commands;

means for inserting, replacing or deleting said stored character strings, and

means for updating the stored tokens in response to said entered commands and in response to insertion, replacement, and deletion of stored character strings.

3. An editing system according to claim 1, characterized in that said speech recognizing unit includes means for recognizing individual characters and furnishing respective character coded identifiers, and in that said dictionary and lookup means is arranged to issue single-character character strings in response to character coded identifiers, and in that said editing means has a command entry means for receiving input commands which selectively switch the editing means between a work mode and a spell mode to enable input and editing selectively either for whole words or single characters.

4. An editing system according to claim 1,

having a speech recognizing unit that is capable of recognizing spoken commands and issuing respective coded identifiers,

the system being further characterized in that said editing means includes control means for receiving editing commands in coded form either from the speech recognizing unit or from a keyboard or from a manual input device.

5. An editing system according to claim 1,

having a speech recognizing unit that issues a special condition indication when a work is not clearly recognized,

said editing means in response to such special condition indication, optically identifying, on the display means, text which the speech recognizing unit does not clearly recognize.

6. An editing system according to claim 1, in which the display means has a cursor function for optically highlighting a stored character string being edited, the system further comprising

means for selectively varying the length of the displayed cursor, the length being determined in response to the character count value in the token associated with the character string being edited.

7. A method of editing spoken input represented in a visual form, in a system comprising a speech recognizing unit furnishing coded identifiers of recognized entities such as spoken words or spoken punctuation marks, the method comprising the step of:

for each said coded identifier of an entity to be represented by a word character string or a punctuation character, furnishing a respective token associated with said each coded identifier, each token indicating whether the work character string or punctuation character is to be concatenated by a space at either or both ends.

8. A method as in claim 7 comprising the further step of:

indicating in each token the number of characters in the character string representing the coded identifier associated with the token, and

in accordance with predefined context rules and in response to a sequence of tokens furnished for successive coded identifiers and respective characters representing the recognized entities in a basic form, generating formatted output data representing the spoken input in a text readable format including spacings.

9. A method according to claim 8, including the further steps of:

providing, for each token, editing control information including a capitalization indication;

storing in a storing and formatting element, all tokens of a file of entities including words and punctuation marks being edited, and

in response to inserting, replacing, or deleting a work character string, updating stored tokens to reflect correctly the required concatenation, capitalization, and character count of successive entities of text.

10. A method according to claim 8, including the further step of:

after occurrence of a punctuation mark, updating the token of the next coded identifier to include an indication for capitalization of the coded identifier.

11. A method according to claim 8, further including the step of:

generating a special code work if a spoken work or punctuation mark cannot be recognized by said recognizing unit, and

in response to the generating of said special code word, generating a special character string and an associated token and displaying a dummy word and storing a token therefor for subsequent editing.

12. In a speech recognition system that generates recognition codes in response to spoken word inputs wherein each of at least some of the recognition codes corresponds to a spoken text word or a spoken punctuation work, a text editor comprising:

lookup memory means for assigning a token to each said generated recognition code corresponding to a text word or a punctuation word, wherein each assigned token includes concatenation indicia that indicates how the recognition code to which the token is assigned is permitted to connect to spaces and to other generated recognition codes to which tokens are assigned.

13. In a system as in claim 12 which has a speech recognizing unit which provides a recognition code for each spoken word input, the text editor further comprising:

dictionary storage means which receives each recognition code as an input and ascribes to each recognition code corresponding to a text word or a punctuation word a corresponding string of characters as an output;

each string of characters representing a recognition code to which a token is assigned.

14. In a system as in claim 13 wherein at least some of the code identifiers correspond to editing command spoken inputs, a text editor further comprising:

means for producing a command word in response to a code identifier output corresponding to one of said editing command spoken inputs; and

storing and formatting means which receives the produced command word as input and which, in response to the receipt of the produced command word, enters the text editor into an editing mode corresponding to the command word.

15. A text editor as in claim 14 further comprising:

means for indicating the number of characters in each string of characters corresponding to a spoken text word or a spoken punctuation word.

16. A text editor as in claim 15 wherein the string length indicating means includes said lookup memory wherein the tokens generated by said lookup memory include a character count for each string of characters to which a token is assigned.

17. A text editor as in claim 16 wherein said storing and formatting means includes:

means for producing context information based on text words, punctuation words, and commands previously received as recognition codes; and

means for directing the context information back to said lookup memory and for updating the tokens assigned to each string of characters.

18. A text editor as in claim 17 further comprising:

means for altering a file after said storing and formatting means enters the text editor into one of the editing modes;

wherein at least some of the alterations in the file result in changes in the context information and in changes in the tokens assigned to respective strings of characters.

19. A text editor as in claim 18 wherein said storing and formatting means includes:

means for providing a text formatted communication output capable of being sensed by a user; and

user-controlled means for sending commands to said storing and formatting means;

the user being thereby able to alter the mode of the text editor in response to sensed formatted communication output.

20. A system for recognizing spoken words, the system having editing capabilities and comprising:

a speech recognizing unit which represents each spoken word with a code identifier; and

a text editor including;

dictionary storage means for assigning some code identifiers to commands and some code identifiers to respective character strings of text words; and

lookup means for assigning a respective token to each character string of a text work, each token including concatenation information indicating whether the character string to which said each token is assigned is to have a space or no space therebefore and thereafter.

21. A system as in claim 20 wherein said dictionary storage means assigns some code identifiers to single characters with which words are spelled and wherein the system further comprises:

storing and formatting means for (a) receiving character strings, command, and single characters from said dictionary means and tokens from said lookup means and (b) generating context information therefrom.

22. A system as in claim 21 wherein said text editor further comprises:

means for entering the text editor into a spellmode in which single characters are the spoken inputs;

means for detecting when a code identifier has no command word, character string, or single character corresponding thereto; and

means for storing a new word into said dictionary storage means with a corresponding character string and a corresponding token in said lookup means in response to the spelling of the new word character by character after detection of a code identifier which has no command word, character string, or single character corresponding thereto.