Voice recognition apparatus, method, and computer program product

1. A voice recognition apparatus comprising:

a program information storage unit configured to store metadata of a broadcast program in association with a user's viewing state;

a dictionary storage unit configured to store a recognition dictionary including a recognition word which is a voice recognition target and a priority of the recognition word;

a recommendation state storage unit configured to store a recommendation state representing one of a first state and a second state, the first state representing a state in which a feature word having a high first score is recommended among feature words, each feature word representing a feature of the broadcast program, the first score representing a degree of the user's preference for a feature word, the second state representing a state in which a feature word including specific information is recommended among the feature words, the specific information specifying a broadcast program having a high second score, the second score representing a degree of the user's preference for the broadcast program;

a calculating unit configured to calculate the first score, the second score, and the feature words on the basis of the metadata and the viewing state;

an updating unit configured to update the priority of a recognition word including a feature word among the feature words according to the first score when the recommendation state indicates the first state and to update the priority of a recognition word including the specific information according to the second score when the recommendation state indicates the second state;

a receiving unit configured to receive a voice;

a recognizing unit configured to recognize the voice using the recognition dictionary; and

an operation control unit configured to control an operation on the broadcast program based on a recognition result by the recognizing unit.

2. The apparatus according to claim 1 , further comprising

a display control unit configured to preferentially select a feature word that has a high first score as an input candidate representing a candidate of an input indicating the operation, and causes the selected input candidate to be displayed on a display unit.

3. The apparatus according to claim 2 ,

wherein

the display control unit

preferentially selects a feature word that has a high first score as the input candidate when selection is designated based on the first score,

preferentially selects the feature word that includes the specific information specifying the broadcast program having the high second score as the input candidate when selection is designated based on the second score, and

causes the selected input candidate to be displayed on the display unit.

4. The apparatus according to claim 1 ,

wherein

the updating unit

updates the priority of the recognition word including the feature word according to the first score when update is designated based on the first score, and

updates the priority of the recognition word including the specific information according to the second score when update is designated based on the second score.

5. The apparatus according to claim 1 ,

wherein the calculating unit calculates the first score such that a value on a feature word included in the broadcast program viewed by the user is higher than a value on a feature word included in the broadcast program not viewed by the user.

6. The apparatus according to claim 1 ,

wherein

the program information storage unit stores the metadata in association with the viewing state for each of a plurality of the users,

the dictionary storage unit stores the recognition dictionary for each of the plurality of the users,

the calculating unit calculates the feature word and the first score based on the metadata and the user's viewing state for each of the plurality of the users, and

the recognizing unit further determines the user who has uttered the voice based on the received voice, and recognizes the voice using the recognition dictionary of the determined user.

7. A voice recognition method executed in a voice recognition apparatus that includes a program information storage unit that stores metadata of a broadcast program in association with a user's viewing state, a dictionary storage unit that stores a recognition dictionary including a recognition word which is a voice recognition target and a priority of the recognition word, and a recommendation state storage unit that stores a recommendation state representing one of a first state and a second state, the first state representing a state in which a feature word having a high first score is recommended among feature words, each feature word representing a feature of the broadcast program, the first score representing a degree of the user's preference for a feature word, the second state representing a state in which a feature word including specific information is recommended among the feature words, the specific information specifying a broadcast program having a high second score, the second score representing a degree of the user's preference for the broadcast program, the method comprising:

calculating the first score, the second score, and the feature words on the basis of the metadata and the viewing state;

updating the priority of a recognition word including a feature word among the feature words according to the first score when the recommendation state indicates the first state and updating the priority of a recognition word including the specific information according to the second score when the recommendation state indicates the second state;

receiving a voice;

recognizing the voice using the recognition dictionary; and

controlling an operation on the broadcast program based on a recognition result at the recognizing of the voice.

8. A computer program product comprising a non-transitory computer-readable medium including programmed instructions for voice recognition method executed in a voice recognition apparatus that includes a program information storage unit configured to store metadata of a broadcast program in association with a user's viewing state, a dictionary storage unit configured to store a recognition dictionary including a recognition word which is a voice recognition target and a priority of the recognition word, and a recommendation state storage unit configured to store a recommendation state representing one of a first state and a second state, the first state representing a state in which a feature word having a high first score is recommended among feature words, each feature word representing a feature of the broadcast program, the first score representing a degree of the user's preference for a feature word, the second state representing a state in which a feature word including specific information is recommended among the feature words, the specific information specifying a broadcast program having a high second score, the second score representing a degree of the user's preference for the broadcast program, wherein the instructions, when executed by a computer, cause the computer to perform:

calculating the first score, the second score, and the feature words based on the metadata and the viewing state;

updating the priority of a recognition word including a feature word among the feature words according to the first score when the recommendation state indicates the first state and updating the priority of a recognition word including the specific information according to the second score when the recommendation state indicates the second state;

receiving a voice;

recognizing the voice using the recognition dictionary; and

controlling an operation on the broadcast program based on a recognition result at the recognizing of the voice.