Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1999-03-01
2001-11-06
Dorvil, Richemond (Department: 2741)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S275000
Reexamination Certificate
active
06314398
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention is directed to an apparatus and method using speech understanding for automating the channel selection process in interactive television. More particularly, the present invention is directed to a system which utilizes natural language processing for receiving a spoken channel selection request and processing the request against information extracted from an electronic programming guide for automatically selecting a channel.
2. Discussion
The design of interactive television allows two-way communication. Unlike conventional one-way television (TV) or radio broadcasts, most interactive TV enables viewers to respond by telephone. Interactive TV helps people understand complex programs and absorb large amounts of information quickly. Digital television is a new, more efficient method of television transmission which can broadcast several television channels into the space currently used to carry a single analog channel. This will create opportunities for many more new channels and program services.
In the future, it is expected that all television services will be transmitted digitally. By using digital technology, approximately 200 channels, perhaps more, will be made available. Digital transmission also has the potential to offer other advantages to the viewer, such as high definition or wide-screen pictures, CD-quality sound, and near “video-on-demand,” where a film is shown with different start times on several different channels so that the viewer can choose a convenient time to start watching. Interactive services such as home banking, home shopping and connection to the internet could also be made available digitally through the television set.
Viewers who choose to receive digital television by any of these methods will likely need to either buy or rent a special digital receiver or set-top box decoder which will enable digital television pictures to be reassembled on screen. Television sets will also be available with this decoder built in. The set-top box will include a processor that can be used to provide additional services such as speech recognition and speech understanding.
As the number of television channels increase, the viewer will have more and more difficulty making a channel selection. Instead of finding the desired channel by browsing through the entire listing of channels, channel selection can be made by understanding the semantic content of what the viewer wants to see through analyzing the content of the spoken request. By using natural language and dialogue, the viewer would be able to select the desired program and television channel using spoken requests.
In view of the above, it is desirable to create a knowledge representation of electronic program guide (EPG) information which is broadcasted as one of many television signals, and store this information in the TV or in the set-top box. It is also desirable to provide a system which can extract the semantics of the user's spoken program selection request, and apply this request against the EPG knowledge database for searching for the desired program and switching the television to the corresponding channel. Finally, it is desirable to provide a system for interacting with the user by employing a dialogue phase for clarifying the user request or resolving ambiguities.
SUMMARY OF THE INVENTION
The method and apparatus of the present invention utilizes a speech understanding technique for automatically selecting channels on a television or video recording device after searching the content of the broadcasts and an electronic programming guide (EPG). This technique uses natural language and dialogue for selecting the desired TV program based upon spoken requests. Using the content of the EPG containing TV program information along with summaries of some events, such as movies, a dynamic recognition vocabulary is created. This vocabulary contains the key words and phrases that can be recognized and understood by the system. This vocabulary is completed by non-key words that are likely to be used by a viewer when selecting channels (e.g. “I would” or “please”).
A database of grammar data structures containing a priori knowledge about the meaning of key words such as “sports” or “movies” is used by a natural language processor to understand the semantic content of the spoken request. Furthermore, a representation of the channel selection task semantics is used by a dialogue manager to help the user in performing the request. The natural language processor utilizes local and global parsing modules for identifying the semantically relevant portions of the spoken request and sends this information to the dialogue manager. The natural language processor also organizes the words and phrases, once understood, into a semantic representation of the spoken request.
The channel selection semantic representations contain information such as the program to select, the time, possibly the channel or network, etc. The viewer can select a channel but also ask information about what will be available on a particular day or time. The information can be provided to the user as audio responses or displayed on the television screen. Because of the natural language processor, the viewer can concentrate on his/her goal without worrying about the style of language he/she uses to communicate with the speech understanding device. A split screen mode may be used when there is ambiguity to allow the user to further refine the request.
A history database of the user preferences (in terms of preferred sports or movie types) is built automatically by the dialogue manager. This history database may contain records of past spoken requests and dialogues which can be used by the dialogue manager to dynamically modify the language model of the speech understanding device to favor some of the words during recognition. This history database can also be used during the dialogue phase to rank the questions to ask according to the user preferences.
The speech understanding device can also be used for programming a video recording device by linking the channel selection to the record command of the recording device. For programs that are recorded frequently, macros can be established as an additional feature for the user. Similarly, reminders can be used to let the viewer know, for example, that a basketball game will be broadcasted tomorrow.
REFERENCES:
patent: 4305131 (1981-12-01), Best
patent: 5663757 (1997-09-01), Morales
patent: 5809471 (1998-09-01), Brodsky
patent: 5832439 (1998-11-01), Cox, Jr. et al.
patent: 5844620 (1998-12-01), Coleman et al.
patent: 5900905 (1999-05-01), Shoff et al.
patent: 6133909 (2000-10-01), Schein et al.
Contolini Matteo
Junqua Jean-claude
Dorvil Richemond
Harness & Dickey & Pierce P.L.C.
Matsushita Electric - Industrial Co., Ltd.
LandOfFree
Apparatus and method using speech understanding for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method using speech understanding for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method using speech understanding for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2618600