Abstract: A camera movement amount acquisition unit (101) acquires as a camera movement amount (110), an estimated movement amount of a vehicle (10) calculated using photographed image data acquired by, a camera (20) installed on the vehicle (10). In synchronization with the acquisition of the camera movement amount (110) by the camera movement amount acquisition unit (101), a LiDAR movement amount acquisition unit (102) acquires as a LiDAR movement amount (120), an estimated movement amount of the vehicle (10) with higher estimation accuracy than estimation accuracy of the camera movement amount (110). A comparison decision unit (103) compares the camera movement amount (110) with the LiDAR movement amount (120), and decides to use the camera movement amount (110) for a calibration operation when a difference between the camera movement amount (110) and the LiDAR movement amount (120) is less than a threshold value.

Abstract: A speech separation device (12) of a speech separation system includes a feature amount extraction unit (121) configured to extract time-series data of a speech feature amount of mixed speech, a block division unit (122) configured to divide the time-series data of the speech feature amount into blocks having a certain time width, a speech separation neural network (1b) configured to create time-series data of a mask of each of a plurality of speakers from the time-series data of the speech feature amount divided into blocks, and a speech restoration unit (123) configured to restore the speech data of each of the plurality of speakers from the time-series data of the mask and the time-series data of the speech feature amount of the mixed speech.

Abstract: In an acoustic signal separation device (1), a determination unit (6) determines whether or not components from a plurality of sound sources are mixed in each of acoustic signals of respective components regenerated by a signal regeneration unit (5), and when it is determined that a plurality of components is mixed, a series of processes by a feature value extraction unit (2), a data estimation unit (3), a data classification unit (4), and a signal regeneration unit (5) is repeatedly executed until acoustic signals of the components of the respective sound sources are regenerated.

Abstract: A speech separation device (12) of a speech separation system includes a feature amount extraction unit (121) configured to extract time-series data of a speech feature amount of mixed speech, a block division unit (122) configured to divide the time-series data of the speech feature amount into blocks having a certain time width, a speech separation neural network (1b) configured to create time-series data of a mask of each of a plurality of speakers from the time-series data of the speech feature amount divided into blocks, and a speech restoration unit (123) configured to restore the speech data of each of the plurality of speakers from the time-series data of the mask and the time-series data of the speech feature amount of the mixed speech.

Abstract: In an acoustic signal separation device (1), a determination unit (6) determines whether or not components from a plurality of sound sources are mixed in each of acoustic signals of respective components regenerated by a signal regeneration unit (5), and when it is determined that a plurality of components is mixed, a series of processes by a feature value extraction unit (2), a data estimation unit (3), a data classification unit (4), and a signal regeneration unit (5) is repeatedly executed until acoustic signals of the components of the respective sound sources are regenerated.

Abstract: An information processing device includes processing circuitry to acquire a signal representing voices corresponding to utterances made by one or more users; recognize the voices from the signal, convert the recognized voices into character strings to identify the utterances, and identify times corresponding to the utterances; identify users who have made the utterances, as speakers from among the users; store information including records indicating the utterances, the times corresponding to the utterances, and the speakers corresponding to the utterances; estimate meanings of the utterances; refer to the information and when a last utterance of the utterances and one or more of the utterances immediately preceding the last utterance are not a conversation, determine that the last utterance is a voice command for controlling a target; and when it is determined that the last utterance is the voice command, control the target in accordance with the meaning of the last utterance.

Abstract: An object is to provide a technique which can provide a highly valid recognition result while preventing unnecessary processing. A voice recognition device includes first to third voice recognition units, and a control unit. When it is decided based on recognition results obtained by the first and second voice recognition units to cause the third voice recognition unit to recognize an input voice, the control unit causes the third voice recognition unit to recognize the input voice by using a dictionary including a candidate character string obtained by at least one of the first and second voice recognition units.

Abstract: A detector 5 detects words acoustically similar to each other from text information, and a selector 7 selects a synonym from a storage 6, the synonym corresponding to a word detected by the detector 5 and being acoustically neither the same as nor similar to any word existing in the text information. A replacer 8 replaces the word detected by the detector 5 with the synonym selected by the selector 7.

Abstract: An object is to provide a technique which can provide a highly valid recognition result while preventing unnecessary processing. A voice recognition device includes first to third voice recognition units, and a control unit. When it is decided based on recognition results obtained by the first and second voice recognition units to cause the third voice recognition unit to recognize an input voice, the control unit causes the third voice recognition unit to recognize the input voice by using a dictionary including a candidate character string obtained by at least one of the first and second voice recognition units.

Abstract: A detector 5 detects words acoustically similar to each other from text information, and a selector 7 selects a synonym from a storage 6, the synonym corresponding to a word detected by the detector 5 and being acoustically neither the same as nor similar to any word existing in the text information. A replacer 8 replaces the word detected by the detector 5 with the synonym selected by the selector 7.

Abstract: A search unit 60 calculates a character-string similarity by acquiring expansion source document data partially matching a search query 70 from a character-string search index 40, determines a relational-data expansion similarity by acquiring expansion destination document data defined to have a relationship with the expansion source document data from a relational-data expansion index 50 and correcting the acquired data by the character-string similarity of the expansion destination document data and relation weight data 80, and outputs a search result 90 arranged in an order of the relational-data expansion similarity.

Abstract: An index generating unit divides each name data of search target data both into words and into characters, calculates start and end scores showing a start and an end of each of the words and start and end scores showing a start and an end of each of the characters, links them to each entry word which constructs the name data as a list (a name ID, a position, and start and end scores), and stores this list in an index storage unit. A searching unit decomposes an input character string into partial character strings, acquires corresponding candidate entries from the index storage unit, and judges the continuity between candidate entries on the basis of lists to add a comparison score according to the continuity to a candidate entry.

Abstract: Voice recognition is realized by a pattern matching with a voice pattern model, and when a large number of paraphrased words are required for one facility, such as a name of a hotel or a tourist facility, the pattern matching needs to be performed with the voice pattern models of all the paraphrased words, resulting in an enormous amount of calculation. Further, it is difficult to generate all the paraphrased words, and a large amount of labor is required.

Abstract: A search unit 60 calculates a character-string similarity by acquiring expansion source document data partially matching a search query 70 from a character-string search index 40, determines a relational-data expansion similarity by acquiring expansion destination document data defined to have a relationship with the expansion source document data from a relational-data expansion index 50 and correcting the acquired data by the character-string similarity of the expansion destination document data and relation weight data 80, and outputs a search result 90 arranged in an order of the relational-data expansion similarity.

Abstract: An index generating unit divides each name data of search target data both into words and into characters, calculates start and end scores showing a start and an end of each of the words and start and end scores showing a start and an end of each of the characters, links them to each entry word which constructs the name data as a list (a name ID, a position, and start and end scores), and stores this list in an index storage unit. A searching unit decomposes an input character string into partial character strings, acquires corresponding candidate entries from the index storage unit, and judges the continuity between candidate entries on the basis of lists to add a comparison score according to the continuity to a candidate entry.

Abstract: Voice recognition is realized by a pattern matching with a voice pattern model, and when a large number of paraphrased words are required for one facility, such as a name of a hotel or a tourist facility, the pattern matching needs to be performed with the voice pattern models of all the paraphrased words, resulting in an enormous amount of calculation. Further, it is difficult to generate all the paraphrased words, and a large amount of labor is required.

Abstract: In conventional retrieval techniques for narrowing down a candidate when there are a plurality of candidates as retrieval results, a method results in a huge number of candidates in order to prevent omission of candidates, whereby the administration cost is increased, and another method involves always processing all candidates, whereby the processing time becomes longer and the response performance is lowered.

Abstract: A search device includes a partial character string extracting unit for acquiring partial character strings for search from a search query inputted, a partial character string searching unit for acquiring name text candidates and pieces of partial character string appearance position information respectively showing the appearance positions of the partial character strings within the name text candidates according to the partial character strings for search, a candidate counting unit for counting an accumulated score for each name text candidates by providing consistency among the appearance positions in consideration of the pieces of partial character string appearance position information in such a way that the appearance positions do not overlap one another in each name text candidate, a candidate-to-be-presented selecting unit for determining a candidate to be presented according to the accumulated score, and a candidate presentation unit for presenting the candidate to be presented.