专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

51. 发明专利

JP2000039899A SPEECH RECOGNITION APPARATUS 失效
公开(公告)号：JP2000039899A
公开(公告)日：2000-02-08
申请号：JP20748498
申请日：1998-07-23
申请人： HITACHI LTD
发明人： KUJIRAI TOSHIHIRO , HATAOKA NOBUO , AMANO AKIO , ODAKA TOSHIYUKI , MURAMATSU RYUJIRO , MATSUDA TOSHIYUKI , SATO HITOSHI
IPC分类号： G10L15/06 , G10L15/10
摘要： PROBLEM TO BE SOLVED: To enable speech recognition with a high recognition rate by classifying the inputted characteristic quantity vector strings which are previously subjected to buffering according to the vowel systems regarded to be likely by a collation section and utilizing these strings as the time series average of the characteristic quantity vector strings inputted with each of vowels. SOLUTION: A correction vector 904 is subtracted from the characteristic quantity vector string 901 and the corrected characteristic quantity vector string 902 is calculated in a correction vector subtraction section 101. The characteristic quantity vector string 902 is subjected to collation and likelihood calculation in the collation section 201 and a recognition candidate 903 is determined. On the other hand, the characteristic quantity vector string 901 for one utterance stored in a buffer 106 is classified by a clustering section 103 and the average for each vowel is calculated in an average calculation section 104. The difference between the standard pattern environmental characteristic quantity vectors and the average of each of the vowels is a value including the personal characteristics of the vowels and there is a correlation among the personal characteristics of each of the vowels. The speaker characteristics and the environmental characteristics are separated and estimated in a character sepn. estimation section 301 by utilizing this correlation.

52. 发明专利

JPH11248815A METHOD AND SYSTEM FOR PROVIDING POSITION INFORMATION AND FM TRANSMISSION STATION AND PORTABLE TERMINAL APPARATUS USED IN THE SYSTEM 失效
公开(公告)号：JPH11248815A
公开(公告)日：1999-09-17
申请号：JP4733798
申请日：1998-02-27
申请人： HITACHI LTD
发明人： OHIRA EIJI , SUKETA HIROKO , HATAOKA NOBUO
IPC分类号： G01C21/00 , G01C21/28 , G01S5/02 , G01S19/05 , G01S19/07 , G01S19/34 , G08G1/005 , H04H1/00 , H04H40/45 , H04W84/10 , H04W88/02 , G01S5/14 , H04H5/00 , H04Q7/34
摘要： PROBLEM TO BE SOLVED: To detect a position both indoors and outdoors and make a terminal compact and consume less power by using a portable terminal for the DGPS to which a means for taking out positional information from signals of a sound band demodulated from received FM modulated signals is added. SOLUTION: This system comprises an FM transmission station which generates positional information by PB signals (push button signals) and transmits the PB signals by FM waves of 10 mW power, and a DGPS terminal in which a function of extracting PB signals from FM demodulated sound signals is added to a CPU 1. An output signal digitized at an AD conversion part 7 is taken out inside the CPU 1 as positional information represented by numerals and symbolic characters. When and AD converter is set outside, the signal digitized by the AD conversion part 7 is taken into the CPU 1 via a bus 2. Accordingly, a position whether it is indoors or outdoors can be detected and the terminal can be made compact and to consume less power.

53. 发明专利

JPH11202886A SPEECH RECOGNITION DEVICE, WORD RECOGNITION DEVICE, WORD RECOGNITION METHOD, AND STORAGE MEDIUM RECORDED WITH WORD RECOGNITION PROGRAM 失效
公开(公告)号：JPH11202886A
公开(公告)日：1999-07-30
申请号：JP449098
申请日：1998-01-13
申请人： HITACHI LTD
发明人： OBUCHI YASUNARI , OBA SHINYA , KONDO KAZUO , WAKIZAKA SHINJI , TOUSHITA TETSUJI , HATAOKA NOBUO , AMANO AKIO , ODAKA TOSHIYUKI , KOKUBO HIROAKI
IPC分类号： G10L15/06 , G10L3/00
摘要： PROBLEM TO BE SOLVED: To improve the recognizing performance of the whole system by estimating the recognizing performance in a state wherein the contents of a dictionary are modified and previously deciding whether or not estimated modification is accompanied by deterioration in the performance of the system. SOLUTION: A user inputs modification contents of the dictionary such as a word to be added and a word to substitute. A recognizing performance decision part 104 estimates how the recognizing performance of the whole system changes when the modification that the user has inputted is actually done. A conversion candidate generation part 106 regards the modification contents that the user has specified as a candidate word or generates a candidate similar to the modification contents that the user has specified according to the decision result and outputs it on condition that set end reference is satisfied. When the user approves the output, a dictionary update pad 108 updates the contents of the word dictionary 112. Further, a dictionary update hypothesis generation part 110 generates hypotheses for the removal of an arbitrary word from the word dictionary 112 at all times.

54. 发明专利

JPH11102280A SOUND OUTPUT METHOD FOR IMAGE INFORMATION 失效
公开(公告)号：JPH11102280A
公开(公告)日：1999-04-13
申请号：JP20895598
申请日：1998-07-24
申请人： HITACHI LTD
发明人： ARITSUKA TOSHIYUKI , HATAOKA NOBUO , NENE YOSHITO , TAKAHASHI JUICHI
IPC分类号： G10L13/00 , G06F3/033 , G06F3/048 , G06F3/16 , G06T11/60 , G06T11/80 , G09B21/00 , G10L21/06 , G10L3/00
摘要： PROBLEM TO BE SOLVED: To enable a visually handicapped person who has the difficulty in making good use of a GUI to use GUI adapted equipment by controlling the output time of a sound allocated so as to indicate the shape of an image object and controlling the gradation of the representative color of the image object so that it corresponds to the variation of the output sound, and thus auditorily outputting the shape of the image object as sound information. SOLUTION: For example, the lengths of the respective sides of an object are (a), (b), (c), and (d). Therefore, the length L of the outline of the object in a drawing space is represented as L=a+b+c+d. Then the sound allocated to the object in advance is outputted for a time prescribed by L once the cursor enters the object. A sound which is different in pitch and tune depending upon shape of the object is selected as the sound assigned to the sound representation of the object. Namely, the specific sound is made to correspond to the characteristics of the object in advance and outputted for the time length prescribed by the length L of the outline of the object and the positions of the vertexes to make it possible to understand the shape of the object from the sound.

55. 发明专利

JPH1188507A SPEECH RECOGNITION SYSTEM FOR PAGER 失效
公开(公告)号：JPH1188507A
公开(公告)日：1999-03-30
申请号：JP24276997
申请日：1997-09-08
申请人： HITACHI LTD
发明人： MATSUDA TOSHIYUKI , SATO HITOSHI , MURAMATSU RYUJIRO , NAKAYAMA TAKURO , KOSAKA MASAHIRO , SUZUKI TOSHIAKI , KUJIRAI TOSHIHIRO , HATAOKA NOBUO , MATSUHASHI SATOSHI
IPC分类号： G10L15/00 , G10L15/22 , H04M3/42 , G10L3/00
摘要： PROBLEM TO BE SOLVED: To simplify the input operation and to reduce the input time by accepting a correction request of the user only on the occurrence of a mistake by speech recognition and correcting the input to attain a smooth input. SOLUTION: The arrival of a call from a public network 20 is reported to a recognition control section 32 via a line interface 31, the call is connected to a voice interactive processing section 40 after the notice and connection to the public network 20 is instructed to the interface 31 on the other hand. Through the connection and the interaction with the user, contents of a message are confirmed with the user, and if required, the contents are corrected on the user request, and the contents of the correct message and a message destination pager number are stored in the recognition control section 32. After the end of interaction, the voice interactive processing section 40 instructs line interruption to the interface 31. The recognition control section 32 sends the message simply and quickly the destination pager terminal, based on the received information. The recognition control section 32 always monitors the operating state of the interface 31 and the voice interactive processing section 40.

56. 发明专利

JPH1152978A ADAPTIVE TYPE VOICE RECOGNIZING DEVICE 失效
公开(公告)号：JPH1152978A
公开(公告)日：1999-02-26
申请号：JP20577597
申请日：1997-07-31
申请人： HITACHI LTD
发明人： OBUCHI YASUNARI , AMANO AKIO , ODAKA TOSHIYUKI , HATAOKA NOBUO
IPC分类号： G10L15/06 , G10L15/10 , G10L3/00
摘要： PROBLEM TO BE SOLVED: To enable making an acoustic model having high accuracy by learning mutual relation between correction quantity of an acoustic models, and holding the result as previous knowledge. SOLUTION: A previous learning section 102 sends input of a teacher signal 108 and voice data 110 for previous learning to a after adaptation model making section 112. After that, the after adaptation model making section 112 compares a before adaptation acoustic model 114 with an after adaptation acoustic model 126, and outputs an acoustic model correction quantity 115. The acoustic model correction quantity 115 is sent to a previous knowledge extracting section 116, previous knowledge is extracted, and held in a previous knowledge holding section 118. Voice data 122 for adaptation collected in environment in which a recognizing device is actually used is inputted to an acoustic model adaptation section 104. Inputted voice data 129 for recognition is sent to a recognition executing section 130 with a word dictionary 128 for recognition in a voice recognizing section 106, these data are compared and collated with the after adaptation acoustic model 126, and a recognized result 132 is outputted.

57. 发明专利

JPH10336124A INFORMATION PROVIDING DEVICE FOR PORTABLE TERMINAL 失效
公开(公告)号：JPH10336124A
公开(公告)日：1998-12-18
申请号：JP13965997
申请日：1997-05-29
申请人： HITACHI LTD
发明人： OHIRA EIJI , SUKETA HIROKO , HATAOKA NOBUO
IPC分类号： H04H20/00 , H04H20/02 , H04H20/28 , H04H20/61 , H04H60/07 , H04H1/00
摘要： PROBLEM TO BE SOLVED: To obtain an information providing device for a portable terminal which quickly retrieves information that is suitable for user's service environment (time, place and purpose) without increasing the scale of a portable terminal for information receiving. SOLUTION: An information providing device 1 is connected by an information server 4 and a bidirectional communicating means 5 such as a telephone and receives information that is peculiar to a region around the lace where the device 1 is installed form the device 1 through the means 5 and stores the information. The device 1 has a means to receive FM multiplexed broadcasting, edits received information from the FM multiplexed broadcasting and regional information stored in the storing means, and transmits edited information to a portable terminal 2 in the FM multiplexed broadcasting.

58. 发明专利

JPH09307583A AGENT COMMUNICATION METHOD AND AGENT COMMUNICATION SYSTEM 失效
公开(公告)号：JPH09307583A
公开(公告)日：1997-11-28
申请号：JP11661296
申请日：1996-05-10
申请人： HITACHI LTD
发明人： HATAOKA NOBUO , KIKUCHI HIDEAKI
IPC分类号： G06F15/16 , G06F9/46 , G06F13/00 , G06F15/163 , H04L12/54 , H04L12/58
摘要： PROBLEM TO BE SOLVED: To attain communication with a high efficiency among a plurality of computers by providing a common memory storing data relating to internal processing of the computers to each of a plurality of the computers. SOLUTION: When a start command is given to a server system 500, application agents 5051-505N1 are started depending on the number of databases. The agents access corresponding databases. Each agent in this case reads required information from a blackboard 566 being a common memory and writes information such as a work state sill in progress or a retrieval result onto the blackboard 506. Similarly in a server system 600, application agents 6051-605N2 corresponding to databases 6021-602N2 are started. Each agent reads/writes data from/onto a blackboard 606. On the other hand, a super-agent 504 conducts transmission/management of information between the blackboards 506 and 606.

59. 发明专利

JPH08221245A INFORMATION PROCESSOR RECEIVING VOICE OPERATION INSTRUCTION 失效
公开(公告)号：JPH08221245A
公开(公告)日：1996-08-30
申请号：JP2314595
申请日：1995-02-10
申请人： HITACHI LTD
发明人： ANDOU HARU , MATSUDA YASUMASA , OSUJI SHIGETO , HASEGAWA TSUKASA , KIKUCHI HIDEAKI , HATAOKA NOBUO
IPC分类号： G06F3/16 , G06F17/28 , G10L15/00 , G10L15/10 , G10L15/18 , G10L15/28 , G10L3/00 , G10L9/00
摘要： PURPOSE: To easily execute the fine adjustment on an instructed operation by receiving the adverbial expression showing the degree of an operation together. CONSTITUTION: A voice recognition program 13 has a voice feature amount extraction program 135 and a changed amount calculation program 136. The voice feature extraction program 135 extracts power and pitch from the voice information corresponding to the adverb such as 'a little more', etc., showing the degree of the operation such as the moving amount, etc., that a user utters when the editing operation such as the moving of an object, etc., is instructed. The changed amount calculation program 136 calculates the changed amount for adjusting the degree of the operation such as moving amount, etc., from the extracted power and pitch.

60. 发明专利

JPH08221095A METHOD FOR READING SENTENCE ALOUD 失效
公开(公告)号：JPH08221095A
公开(公告)日：1996-08-30
申请号：JP2307095
申请日：1995-02-10
申请人： HITACHI LTD
发明人： HASEGAWA TSUKASA , ASUMA HAJIME , MATSUDA YASUMASA , HATAOKA NOBUO , ANDOU HARU , KIKUCHI HIDEAKI
IPC分类号： G06F3/16 , G06F17/21 , G10L13/04 , G10L13/06 , G10L5/04 , G10L5/02
摘要： PURPOSE: To read a parenthesis in a sentence and the sentence in parentheses aloud so that a worker (user) understands the contents of the sentence easily by constituting so that the read aloud excepting the parenthesis in the sentence and an annotation sentence, etc., put in by quotation marks is performed. CONSTITUTION: An extraction part 104 detects the data coinciding with the parenthesis and quotation marks data stored in a parentheses/quotation marks memory 105 from the sentence data stored in a sentence data memory 103, and extracts the sentence data excepting the sentence data put in by the detected data to store them in a read aloud sentence data memory 106. A read data conversion part 107 identifies a kanji-kana mixed word and an english word incorporated in the data from the storing read aloud sentence data, and converts them into the read data to store them in a read data memory 108. A voice synthesis part 109 converts the storing read data into the speaking data consisting of the phoneme data, intonation, punctuation, etc., showing a voice, and a voice output part 110 generates an analog signal from the speaking data to output as the voice.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式