JP7202938B2 - Program name search support device and program name search support method - Google Patents

Description

The present embodiment relates to a program name search support device and a program name search support method.

2. Description of the Related Art AV equipment such as televisions generally uses an EPG (Electronic Program Guide) obtained from broadcast waves to search for broadcast programs. When searching for a broadcast program name using a search keyword, conventionally, characters to be searched for have been entered by operating buttons on a remote controller.

In recent years, a voice recognition technology has been put into practical use for recognizing an operator's voice and operating an AV device or the like based on the recognition result (see, for example, Patent Document 1). For example, when an operator utters a search keyword toward a remote control equipped with a microphone, it is possible to search for broadcast program names based on voice recognition results. Such a voice search enables a search to be performed in a short period of time with less trouble than a character input search, and greatly contributes to the improvement of operator's convenience.

However, the voice search has a problem that the program cannot be searched if the program name (character string) obtained as a voice recognition result does not match the program name obtained from the EPG.

JP 2012-168349 A

An object of the present embodiment is to provide a program name search support device and a program name search support method capable of improving accuracy of program name search by voice search.

A program name search support device of the present embodiment includes a program information storage unit that stores a first program name, which is a notation of a program name stored in a program name search target device, and a reading of the program name in pairs; a character data acquisition circuit for acquiring a second program name, which is character data obtained by performing voice recognition processing on audio data based on the pronunciation of the program name stored in the program information storage unit; and a translation dictionary for storing the first program name and the second program name in pairs when the program name and the second program name are different.

1 is a schematic diagram showing an example of configuration of a program name search system using a program name search support device according to the present embodiment; FIG. 1 is a block diagram showing the configuration of a language processing server, which is an example of a program name search support device according to an embodiment; FIG. The figure explaining an example of new program information data. The figure explaining an example of program name reading data. 4 is a flowchart for explaining an example of a method for creating a replacement dictionary;

Embodiments will be described below with reference to the drawings.

FIG. 1 is a schematic diagram showing an example of the configuration of a program name search system using a program name search support device according to this embodiment. The program name search system is composed of a language processing server 1, a broadcast receiver 2 as a program name searched device, and a speech recognition server 3. FIG. The language processing server 1 , the broadcast receiving device 2 and the speech recognition server 3 are connected to each other via the Internet line 4 .

The broadcast receiver 2 receives and reproduces programs provided by broadcasters as radio waves propagating in space, and programs provided by distributors through networks such as cable networks and IP networks. The broadcast receiving apparatus 2 receives operation instructions from a user via a remote controller (hereinafter referred to as a remote controller) 21 . The broadcast receiving device 2 may be configured to include a recording/playback device for recording the received program.

The remote control 21 has operation keys (number keys, arrow keys, color buttons, etc.) and a microphone. For example, by performing a predetermined operation on the remote controller 21, such as pressing a microphone button, which is one of the operation keys, the user can transmit audio data to the broadcast receiver 2 via the microphone of the remote controller 21. can. That is, the user can input an operation instruction to the broadcast receiving device 2 using voice data.

The speech recognition server 3 is a server that provides a cloud-based speech recognition service. The voice recognition server 3 converts voice data transmitted from a device connected to the Internet line 4 into character data and outputs the character data.

A language processing server 1 , which is a program name search support device, has a processor 11 . The language processing server 1 analyzes character data input from the broadcast receiving device 2 or the speech recognition server 3, extracts operation details and parameters for the broadcast receiving device 2, and extracts a format that can be processed by the broadcast receiving device 2. converted to and output.

FIG. 2 is a block diagram showing the configuration of the language processing server 1. As shown in FIG. The language processing server 1 includes a character data communication circuit 51, a language data analysis circuit 52, a command output circuit 53, a reading dictionary registration circuit 54, a voice data transmission circuit 55, and a voice conversion circuit 56. The language processing server 1 also has storage devices for a new program information database 61 and a translation dictionary 62 .

Character data communication circuit 51, language data analysis circuit 52, command output circuit 53, reading dictionary registration circuit 54, voice data transmission circuit 55, and voice conversion circuit 56 are provided by a CPU (Central Processing Unit), which is processor 11. ), or hardware using an FPGA or the like.

A character data communication circuit 51 as a character data acquisition circuit controls transmission and reception of character data with devices connected to the Internet line 4 (for example, the broadcast receiver 2 and the voice recognition server 3). For example, it acquires character data output from the voice recognition server 3, or acquires electronic program guide (EPG) information transmitted from a broadcasting station or the like. Further, when the character data communication circuit 51 acquires new program information from another server (not shown), the character data communication circuit 51 registers the information in the new program information database 61 . Note that the new program information is information of a program having a new program name that has not been acquired so far in the EPG information.

FIG. 3 is a diagram for explaining an example of new program information data. The new program information data shown in FIG. 3 is registered in the new program information database 61. FIG. As shown in FIG. 3, one record is created for each program in the new program information data. and “(Information) Date of Acquisition”. Data extracted from the new program information acquired by the character data communication circuit 51 is set in the "program name" and "reading". In the "acquisition date", the date on which the new program information of the record was received from another server is set.

The language data analysis circuit 52 performs natural language analysis processing such as morphological analysis and syntactic analysis on the character data (text constructed as natural sentences) acquired by the character data communication circuit 51 as necessary, and converts the character data into Understand the meaning (operation content) of For example, when character data "I want to see OO (program name)" is input, it is analyzed as meaning to "search" for "OO (program name)". If a program name is included in the analysis result, a search is made to see if the program name is registered in the replacement dictionary 62 . If it is registered, the character data of the program name obtained as the analysis result is replaced with another character data specified in the replacement dictionary 62 and output.

The command output circuit 53 converts the analysis result of the language data analysis circuit 52 into a format that can be processed by the broadcast receiver 2 and outputs the result. For example, in the case of the above example, an operation instruction signal is output to the broadcast receiving device 2 so as to execute a "search" operation using "00 (program name)" as a keyword.

The voice conversion circuit 56 converts input character data into voice data. The converted voice data is output to voice data transmission circuit 55 .

The voice data transmission circuit 55 transmits voice data input from the voice conversion circuit 56 to a device (for example, the voice recognition server 3) connected to the Internet line 4. FIG.

The reading dictionary registration circuit 54 creates program name reading data and registers it in the reading dictionary 62 . FIG. 4 is a diagram illustrating an example of program name replacement data. As shown in FIG. 4, one record is created for each program in the program name replacement data, and each record has three items, for example, "input program name", "replacement program name", and "registration date". consists of The "input program name" is an item corresponding to the program name (character data) obtained by converting the "pronunciation" of the "program name" in the new program information into character data by the voice recognition server 3. FIG. In the "replacement program name", another character data (specifically, the "program name" of the new program information is set to be used in place of the character data registered in the "input program name"). ” is set with the date when the record was registered in the replacement dictionary 62. The replacement dictionary 62 is referred to by the language data analysis circuit 52. FIG.

Next, a method of creating a replacement dictionary in the language processing server 1 will be described with reference to FIG. FIG. 5 is a flow chart illustrating an example of a method for creating a replacement dictionary.

First, in the character data communication circuit 51, new program information is acquired from an external server or the like, and the information is registered in the new program information database 61 (S1). Subsequently, the reading dictionary registration circuit 54 extracts program names that are candidates for reading replacement dictionary registration from the new program information database 61 (S2). For example, it extracts programs registered in the new program information database 61 after the last day when the replacement dictionary was created. For example, if the date on which the translation dictionary was last created is November 1, 2018, the records whose “acquisition date” is after November 2, 2018 are extracted, and the extracted records have “program name ” is set as a candidate for registration of the replacement dictionary.

It should be noted that the method of extracting candidate programs for registration in the replacement dictionary is not limited to the method described above. For example, each record of the new program information database 61 is set with a flag that can identify whether or not it has been extracted as a replacement dictionary registration candidate. , may be used as a replacement dictionary registration candidate.

Next, the translation dictionary registration circuit 54 converts the program name extracted in S2 into voice data and outputs the voice data to the voice recognition server 3 (S3). Specifically, first, the replacement dictionary registration circuit 54 outputs the extracted program name (character data) to the voice conversion circuit 56 . The voice conversion circuit 56 converts the input character data into voice data and outputs the voice data to the voice recognition server 3 via the voice data transmission circuit 55 . In addition, when a plurality of program names are extracted in S2, one program name is selected from the plurality of program names, and the procedure of S3 described above is executed.

It should be noted that the conversion of the extracted program name (character data) into voice data is not limited to the voice conversion circuit 56, but is performed by another server capable of receiving character data from the language processing server 1 and having a voice conversion function. etc. For example, if the broadcast receiver 2 has a voice conversion circuit, the program name is transmitted from the translation dictionary registration circuit 54 to the broadcast receiver 2 via the character data communication circuit 51 and the Internet line 4, and the broadcast receiver 2 The character data may be converted into voice data by a voice conversion circuit. In this case, the converted voice data is output from the broadcast receiver 2 to the voice recognition server 3 via the Internet line 4 .

Subsequently, the character data communication circuit 51 of the language processing server 1 acquires the recognition result of the speech data output in S3, that is, the character data converted from the speech data from the speech recognition server 3 (S4). Character data communication circuit 51 outputs the acquired character data to reading dictionary registration circuit 54 . The replacement dictionary registration circuit 54 compares the input character data with the program name (character data) output in S3 (S5). If the two do not match (S5, NO), the compared character data is registered as a new record in the replacement dictionary 62 (S6). That is, in S3, the program name (character data) output to the speech recognition server 3 is set as the "replacement program name", and in S4, the character data acquired from the speech recognition server 3 is set as the "input program name", and registration work is performed. A record is newly registered in the replacement dictionary 62 in which the date on which the

For example, in S3, the voice data obtained by converting the program name (character data) written as “3rd year K group”, that is, the voice data obtained by converting the reading “sannenkegumi” into voice, is sent to the voice recognition server 3. Output. If the character data "3rd year K class" is input in S4, the two do not match. Accordingly, the pair of character data "3rd year K group" and "3rd year K group" is registered in the reading dictionary 62 in S6 as the topmost record in the table shown in FIG.

On the other hand, in S5, if the input character data matches the program name (character data) output in S3, registration in the translation dictionary 62 is not performed, and the process proceeds to S7.

If there are a plurality of program names that are candidates for registration in the replacement dictionary extracted in S2, and if all the programs have already undergone a series of procedures (S3 to S6) for determining whether or not they need to be registered in the replacement dictionary 62 (S7, YES ), the procedure for creating a replacement dictionary in the language processing server 1 shown in FIG. 5 is terminated. On the other hand, if there is a program name that has not been subjected to a series of procedures for judging necessity of registration in the translation dictionary 62 (S7, NO), proceed to S8, and extract one program name from among the program names that have not been performed. and set as the name of the program to be judged next. An example procedure of S3 to S6 is executed for the set program name.

In this way, by registering the notation of the program name obtained by voice recognition that differs from the notation of the program name in the electronic program guide (EPG) as a translation dictionary, the broadcast receiver 2 can perform voice search. In this case, the search accuracy can be improved.

For example, in the program name search system shown in FIG. 1, voice search for a specific program name is performed as follows. First, the user uses the remote controller 21 of the broadcast receiving apparatus 2 to voice input that he/she wishes to search for a specific program name. For example, when the user wants to search for a program called "BACK STREET KIDS", the user speaks into the microphone of the remote control 21, "Backstreet and Kids".

The broadcast receiving device 2 transmits voice data input by the user (for example, voice data saying “I want to watch the back story and kids”) to the voice recognition server 3 via the Internet line 4 . The voice recognition server 3 converts the input voice data into character data and transmits the character data to the language processing server 1 via the Internet line 4 . For example, voice data "I want to see Backstreet Kids" is converted into text data "I want to see Backstreet Kids" and is transmitted.

Character data input from the speech recognition server 3 is output from the character data communication circuit 51 to the language data analysis circuit 52 . The language data analysis circuit 52 performs natural language analysis processing on input character data, grasps the meaning and content of the character data, and generates an analysis result. For example, when character data such as "I want to see Backstreet Kids" is input, an analysis result is generated in which an operation of "searching" is executed using the program name "Backstreet Kids" as a keyword.

The language data analysis circuit 52 searches whether or not the program name included in the analysis result is registered in the “input program name” of the translation dictionary 62 . If registered, the character data registered in the "replacement program name" of the record is replaced with the program name included in the analysis result. Then, it outputs the analysis result after replacing the program name to the command output circuit 53 .

For example, if there is a program name "Backstreet Kids" in the analysis result, the language data analysis circuit 52 refers to the translation dictionary 62 and determines that the program name (character data) "Backstreet Kids" is the "input program name". Search whether it is registered as If registered, the program name (character data) registered in the "replacement program name" of the record is extracted.

As described above, if the translation dictionary registration work based on the EPG notation has already been performed, there will be a record in which the "input program name" is set to "Backstreet Kids" and the "replacement program name" is set to "BACK STREET KIDS". , are registered in the translation dictionary 62 . Therefore, the language data analysis circuit 52 refers to the translation dictionary 62 and replaces the program name (character string) "Backstreet Kids" in the analysis result with "BACK STREET KIDS". Then, the command output circuit 53 outputs an analysis result indicating that an operation "search" is executed using the program name "BACK STREET KIDS" as a keyword.

The command output circuit 53 converts the analysis result input from the language data analysis circuit 52 into a format that can be processed by the broadcast receiver 2 and outputs the result. In the case of the above example, an operation instruction signal is output to the broadcast receiving apparatus 2 to "search" for "BACK STREET KIDS" in program names such as a program table (EPG) and recorded data.

The broadcast receiver 2 extracts the program name written as "BACK STREET KIDS" from the stored program table (EPG) or the like, and displays it on the search result screen or the like. When the broadcast receiving device 2 is instructed to search using the notation "Backstreet Kids" which is character data output from the voice recognition server 3, the notation ("BACK STREET KIDS") in the program guide (EPG) is displayed. ), there is a problem that the desired program cannot be searched. On the other hand, in the present embodiment, in the language processing server 1, program names whose notation in the character data output from the speech recognition server 3 differs from the notation in the EPG are replaced with the notation in the EPG, and the broadcast receiving device 2 , it is possible to reduce omissions in search due to inconsistency in notation, and to improve search accuracy in voice search.

In the above description, the case of performing voice search for a program name in the broadcast receiving apparatus 2 has been described as an example. be.

While several embodiments of the invention have been described, these embodiments have been presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and equivalents thereof.

1... language processing server,
2 ... broadcast receiving device,
3 ... voice recognition server,
4 Internet line,
21 ... remote control,
51 ... Character data communication circuit,
52... language data analysis circuit,
53 ... command output circuit,
54 ... reading dictionary registration circuit,
55 ... audio data transmission circuit,
56 ... voice conversion circuit,
61 ... new program information database,
62 ... reading dictionary,

Claims (5)

a program information storage unit for storing a pair of a first program name, which is a notation of a program name stored in a program name search target device, and the reading of the program name;
a character data acquisition circuit for acquiring a second program name, which is character data obtained by performing voice recognition processing on audio data based on the pronunciation of the program name stored in the program information storage unit;
a translation dictionary for storing pairs of the first program name and the second program name when the first program name and the second program name are different;
A program name search support device having 2. The program name search support device according to claim 1, wherein said program name stored in said program name search target device is a program name described in an electronic program guide. further comprising a language data analysis circuit for executing natural language analysis processing of character data obtained by performing voice recognition processing on operation voice data input for operating the device to be searched for program name; When the second program name exists in the character data to be subjected to the natural language analysis processing and the second program name is stored in the replacement dictionary, the data analysis circuit performs the second program name 3. The program name search support according to claim 1 or 2, wherein the natural language analysis process is performed by replacing the first program name stored in the replacement dictionary with the first program name stored in pairs in the replacement dictionary. Device. acquiring a pair of a first program name, which is a notation of a program name stored in a program name search target device, and the reading of the program name;
Acquiring a second program name, which is character data obtained by performing voice recognition processing on voice data based on the reading of the program name,
A program name search support method, wherein when the first program name and the second program name are different, the first program name and the second program name are paired and registered in a translation dictionary. 5. The program name search support method according to claim 4, wherein said program name stored in said program name search target device is a program name described in an electronic program guide.

Images