Text-processing system

• US 5,031,113 A
• Filed: 10/17/1989
• Issued: 07/09/1991
• Est. Priority Date: 10/25/1988
• Status: Expired due to Term

• Current Assignee: Nuance Communications Austria Gmbh (Microsoft Corporation)
• Original Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Inventors: Hollerbauer, Werner
• Primary Examiner(s): KEMENY, EMANUEL
• Application Number:
  US07/422,474
• Time in Patent Office: 630 Days
• Field of Search: 381/45, 381/44, 381/48, 369/29, 364/513.5
• US Class Current: 704/235
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A text-processing system in which the sound signal corresponding to a text is applied not only to a speech-recognition device (1) but also to a first further memory (15) for storage, start marks and end marks being allocated to the words detected by the speech-recognition device (1). These marks are related in time to the sound signals stored in the first further memory (15) and are stored in a second further memory (16). In this way a correlation is obtained between the sound signals stored in the first further memory and the individual words detected by the speech-recognition device. The detected words can be reproduced on a display device (13) thereby enabling the sound signal stored in the first further memory and corresponding to a specific detected word to be reproduced acoustically so as to verify the accuracy of a word detected by the speech-recognition device and to correct an erroneously detected word.

Claims

1. A text-processing system comprising:

means for applying sound signals corresponding to a word text to a speech-recognition device, said speech recognition device including means for detecting the words in the sound signals, means for storing said words in a word memory in the form of digital signals and for supplying said words from said memory to a further processing means, means for reproducing said words as a text by means of a display device and for correcting said words when necessary, a printer for printing said words, means for storing the sound signals in a first further memory, means for allocating start marks and end marks to the words detected by the speech-recognition device under control of a computer which cooperates with a keyboard, which marks are related in time to the sound signals stored in the first further memory, said start marks and end marks being stored in a second further memory, wherein, by means of the start and end marks of a detected word stored in the second further memory, the computer is adapted to establish a correlation with the sound signal corresponding to a detected word and stored in the first further memory, means, under control of the computer, for high-lighting a detected word on the display device and at the same time reproducing the associated audio signal acoustically, and wherein, by means of the keyboard, data is loaded into the computer and a detected word, which is high-lighted on the display device, is corrected, if necessary.

2. A system as claimed in claim 1, wherein the speech recognition device cooperates with a learning device, the speech-recognition device includes means for deriving features characteristic of the sound signal enunciated by a specific speaker of a word from said sound signal, wherein the learning device includes means for analyzing said features by comparison with speech-related parameters applied to said learning device and the word corresponding to the relevant sound signal applied to said learning device and are stored in the speech-recognition device as speaker-related parameters, which can be utilized by the speech-recognition device in the process of speech recognition, means for applying a detected word, which has been high-lighted on the display device and corrected by means of the keyboard, to the learning device for analysis together with the features derived from the the sound signal of the original word, and means for storing in the speech-recognition device speaker-related parameters newly derived in the process of said analysis by the learning device.

3. A system as claimed in claim 1 further comprising a learning device which cooperates with the speech-recognition device which includes means for deriving features which are characteristic of a sound signal enunciated by a specific speaker of a word from said sound signal, wherein the learning device includes means for analyzing said features by comparison with speech-related parameters applied to said learning device and the word corresponding to the relevant sound signal applied to said learning device and are stored in the speech-recognition device, as speaker-related parameters, which are utilized by the speech-recognition device in the process of speech recognition, wherein after all words detected erroneously by the speech-recognition device have been corrected by means of the keyboard, the system further includes means for applying the words forming the complete text together with the features derived from the sound signals of the original words to the learning device for analysis, and means for storing in the speech-recognition device the speaker-related parameters newly derived in the process of said analysis by the learning device.

4. A text processing system comprising:

an input for receiving electric sound signals corresponding to a word text,

a first memory coupled to said input for storing the received sound signals,

a speech-recognition device having an input coupled to the sound signal input and an output coupled to a second memory, means for detecting words in the sound signals, and means for storing said words in a word memory,

a display device for reproducing said words as a text,

a computer coupled to said first memory, to said speech-recognition device and to said display device for controlling the operation thereof,

a keyboard coupled to said computer,

means coupling the second memory to the computer whereby, under control of the computer, said second memory stores start marks and end marks correlated in time to words detected by the speech-recognition device and to sound signals stored in the first memory.

wherein by means of said stored start marks and end marks the computer is operative to establish a correlation with a sound signal corresponding to a detected word and stored in said first memory,

means controlled by the computer for accentuating a detected word displayed on the display device and simultaneously acoustically reproducing the corresponding sound signal stored in the first memory, and wherein

the keyboard is adapted to corrected an accentuated word displayed on the display device when said displayed accentuated words does not match the corresponding acoustically reproduced sound signal.