Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1998-12-21
2001-06-26
Korzuch, William R. (Department: 2741)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S270000, C704S275000, C704S010000
Reexamination Certificate
active
06253176
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention relates to a product including at least one processing unit that can be driven to an active state, for which at least one operating state can be activated, and including a speech recognition device for controlling at least one processing unit as regards its at least one operating state while use is made of a command lexicon containing commands and depending as regards its contents on at least one activated operating state of at least one processing unit in its active state, and which recognition device includes a command lexicon memory in which commands of the command lexicon can be stored in the form of command information and identification information assigned to this command information, and which includes speech recognition means to which can be delivered at least speech information representing a command spoken by a speaker and which are arranged for detecting command information to be assigned to such speech information while use is made of identification information stored in the command lexicon memory and which speech recognition means are arranged for transferring detected command information to a processing unit while a processing-unit operating state corresponding to the spoken command can be activated by detected command information delivered to this processing unit.
The invention further relates to a speech recognition device which is arranged for cooperating with at least one processing unit that can be driven to an active state, for which at least one operating state can be activated, and which is arranged for controlling at least one processing unit as regards its at least one operating state while using a command lexicon comprising commands which depends as regards its contents on at least one activated operating state of at least one processing unit in its active state, and which speech recognition device includes a command lexicon memory in which commands of the command lexicon can be stored in the form of command information and identification information assigned to this command information, and which includes speech recognition means to which can be delivered at least speech information representing a command spoken by a speaker and which are arranged for detecting command information to be assigned to such speech information while use is made of identification information stored in the command lexicon memory and which speech recognition means are arranged for transferring detected command information to a processing unit while a processing-unit operating state corresponding to the spoken command can be activated by detected command information delivered to this processing unit.
The invention further relates to a method of generating a command lexicon for a speech recognition device, which speech recognition device is arranged for controlling at least one processing unit as regards its at least one operating state which processing unit can be driven to an active state—in which at least one operating state can be activated—while use is made of the command lexicon, the command lexicon containing commands which can be stored in the form of command information and identification information assigned to the command information, and the command information and the assigned identification information being stored in a command lexicon memory, and at least speech information spoken by a speaker and representing a command being delivered to speech recognition means, and the speech recognition means while using the identification information stored in the command lexicon memory detecting command information for such speech information and which detected command information can be delivered to a processing unit for activating the operating state of the processing unit corresponding to a spoken command.
Such a product and such a speech recognition device and such a method are known, for example, from the document EP 0 607 615 A1. In the known product the speech recognition device is arranged for cooperating with a plurality of processing units and is used for inputting text elements for building texts and for inputting commands for controlling a processing unit with respect to its operating states. Such a processing unit may be formed, for example, by a text processor implementing the Word program for Windows® or the Wordperfect® program or by a table processor implementing the Excels® program. To the speech recognition device can be connected a microphone by which speech information spoken by a user of the speech recognition device can be inputted in a command input mode as a command or in a text input mode as a text element for a processing unit. The speech recognition device includes speech recognition means which include a command lexicon memory in which the commands forming a command lexicon can be stored, that is, in the form of command information for controlling a processing unit and identification information assigned to this command information, which identification information is formed by socalled phonemes. Speech information inputted as a command by a user is processed in the speech recognition means and compared in the command input mode with identification information stored in the command lexicon memory, command information assigned to the speech information input as a command being generated and delivered to a processing unit for activating an operating state of the processing unit which state corresponds to the command.
Before the known speech recognition device is taken into operation for controlling a processing unit, the speech recognition device is to be adapted to this processing unit in a training mode and voice command processing means are to be added to this processing unit. In the training mode the known speech recognition device can be adapted to different processing units and, additionally, to different operating states of respective processing units so as to achieve a better recognition result of the speech recognition means. Therefore, in the training processing mode a user of the speech recognition device selects commands for an operating state of a processing unit, which commands can be used in this operating state and are suitable for changing the operating state of the processing unit. These commands are stored as command information in a collecting memory of command lexicons assigned to this operating state of this processing unit. Subsequently, the user can speak these commands as speech information into the microphone in the training mode and train the speech recognition means for these commands while identification information is generated which is stored in the collecting memory of command lexicons under the command information. All the command lexicons generated in the manner indicated for each operating state are stored in the collecting memory of command lexicons of the speech recognition device and assigned to the processing units and their operating states in a processing unit table.
When the operating state of a processing unit is changed—which was prepared for cooperating with the speech recognition device by an addition of speech command processing means and for which the known speech recognition device was prepared in the training mode—the processing unit sends information to the speech recognition device about what operating state the processing unit is in. While using the processing unit table, the speech recognition device then detects the command lexicon suitable for the present operating state, generated in combination with all the other command lexicons in the training mode and re-stores this command lexicon of the collecting memory of command lexicons in the command lexicon memory. When a user speaks a command for changing the operating state of the processing unit as speech information into the microphone, the result of the recognition of the spoken command is relatively high, because only a very small number of commands which the user selected for this operating state are used for speech recognition. Command information detected by the speech recognition means assigne
Grabherr Manfred
Janek Gabor
Wutte Heribert
Chawan Vijay B.
Korzuch William R.
Piotrowski Daniel J.
U.S. Philips Corporation
LandOfFree
Product including a speech recognition device and method of... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Product including a speech recognition device and method of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Product including a speech recognition device and method of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2542459