专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

JP2002358778A SEMICONDUCTOR INTEGRATED CIRCUIT DEVICE 失效
公开(公告)号：JP2002358778A
公开(公告)日：2002-12-13
申请号：JP2001162793
申请日：2001-05-30
申请人： HITACHI LTD , NEC CORP
发明人： KOJIMA HIROMITSU , IDE SEIHACHI , ICHIKAWA HIROSHI , SAKAMOTO TATSUYA
IPC分类号： G11C11/403 , G11C11/409
摘要： PROBLEM TO BE SOLVED: To set an optimum over-drive period at the time of auto-refresh operation by varying an over-drive period. SOLUTION: A sense amplifier of a synchronous DRAM is provided with an over-drive control circuit 14. When an auto-refresh signal AREF is inputted at the time of auto-refresh, this over-drive control circuit 14 delays an over-drive control signal ODCT2 by some time by a delay circuit 15, and outputs an over- drive signal SAOD having longer period than that at the time of normal operation. Thereby, overdrive of bit lines can be performed in optimum period even at the time of auto-refresh.

2. 发明专利

JPH0683264A DACTYLOLOGY LEARNING DEVICE 失效
公开(公告)号：JPH0683264A
公开(公告)日：1994-03-25
申请号：JP23562792
申请日：1992-09-03
申请人： HITACHI LTD
发明人： ABE MASAHIRO , SAKO YUTAKA , SAGAWA HIROHIKO , ICHIKAWA HIROSHI , INOUE KIYOSHI , ARAI KIYOSHI , SHIMURA TAKANORI , TODA YUJI
IPC分类号： G09B7/02 , G06Q50/00 , G06Q50/10 , G06Q50/20 , G09B21/00 , G06F15/20
摘要： PURPOSE:To provide a dactylology learning device capable of quickly and accurately outputting dactylology information required by a learner and with a displaying method matched with an individual learner and allowing the individual learner to easily register additional data by himself. CONSTITUTION:A dactylology dictionary 1 stores dactylology data and a dictionary retrieving means 2 retrieves an objective data by an entry word or the retrieval of contents and displays the retrieved data on a display output device 8 by image synthesis in accordance with a learner' request. A dictionary registering means 3 prepares and registers dictionary data by combining an dictylology input, a voice input, a character input, and an expression input. Since required information can easily be registered/retrieved by the device, the learning efficiency of dactylology can be improved.

3. 发明专利

JPH067341B2 失效
公开(公告)号：JPH067341B2
公开(公告)日：1994-01-26
申请号：JP18673383
申请日：1983-10-07
申请人： HITACHI LTD
发明人： ICHIKAWA HIROSHI
IPC分类号： G10L13/00 , G10L13/08 , G10L3/00

4. 发明专利

JPH0556008A VECTOR QUANTIZER 失效
公开(公告)号：JPH0556008A
公开(公告)日：1993-03-05
申请号：JP26735691
申请日：1991-10-16
申请人： HITACHI LTD
发明人： ASAKAWA YOSHIAKI , YAMAZAKI KATSUYA , ICHIKAWA HIROSHI
IPC分类号： G10L19/00 , G10L15/02 , H03M7/30 , H04B14/04
摘要： PURPOSE:To improve the information compression efficiency by storing the result of collation at least by one frame. CONSTITUTION:Upon the receipt of spectrum information 106, an element 431 of each code vector is read from a code book 401 of a collation section 400, and a distance calculation section 402 calculates the distance of vectors 106. In the case of retrieving all numbers, a distance 403 of all code vectors in the code book is calculated and the result is fed to a minimum value detection section 404, and a pointer 405 of a code vector (nearest vector) whose value 403 is minimized is outputted. Thus, when the input vector is expressed by plural code vectors, it is not required to send part of the pointer of the code vector. In this case, since quantization distortion is hardly increased, the information compression efficiency is enhanced.

5. 发明专利

JPH04291398A VOICE RECOGNITION SYSTEM AND VOICE RECOGNITION SYSTEM 失效
公开(公告)号：JPH04291398A
公开(公告)日：1992-10-15
申请号：JP5647491
申请日：1991-03-20
申请人： HITACHI LTD
发明人： ICHIKAWA HIROSHI , AMANO AKIO
IPC分类号： G09B29/10 , G10L11/00
摘要： PURPOSE:To offer the voice recognition system which is suitable for the retrieval of a desired map. CONSTITUTION:An input voice is passed through a voice input part 501 and converted by a voice analytic part 502 into a time series of feature information, which is sent to a collation part 504 through an input pattern buffer 503 and collated with a standard patterns representing respective groups in a standard pattern memory 506. When the respective groups are selected, a control part 508 takes the groups out of an external memory 507 according to the information from a navigation main body 509. As a result of the collation, a decision part 505 selects one or plural groups similar to the input voice and a control part 508 takes standard patterns of words belonging to the selected groups out of an external memory 507 and sends them to the standard pattern memory 506; and the collation part 504 collates the input voice with the standard patterns and the decision part 505 decides and outputs the input voice. A standard pattern conversion part 510 is provided for the size reduction and economization of the memories 506 and 507. Consequently, the high-accuracy, economic voice recognition system which reduces the frequency of voicing as much as possible is realized.

6. 发明专利

JPH0452501B2 失效
公开(公告)号：JPH0452501B2
公开(公告)日：1992-08-24
申请号：JP24198683
申请日：1983-12-23
申请人： HITACHI LTD
发明人： HATAOKA NOBUO , ICHIKAWA HIROSHI , ASAKAWA YOSHIAKI , YAJIMA SHUNICHI
IPC分类号： G06F3/16 , G06F17/21 , G06F17/24 , G10L13/00

7. 发明专利

JPH0454000A VOICE ANALYZING/SYNTHESIZING DEVICE 失效
公开(公告)号：JPH0454000A
公开(公告)日：1992-02-21
申请号：JP16474890
申请日：1990-06-22
申请人： HITACHI LTD
发明人： ASAKAWA YOSHIAKI , ICHIKAWA HIROSHI
IPC分类号： G10L19/06 , G10L19/00
摘要： PURPOSE:To stably obtain the synthetic voice of a high quality by constituting a means for converting a voice spectrum to an impulse response waveform, so as to operate a low-pass component of the voice spectrum. CONSTITUTION:A buffer memory 103 constituted of two faces is provided in order to adjust the time of a processing and to prevent an interruption of an input voice. Also, the device is provided with a frame data segmenting part 104, a DC component adjusting part 106, and a power spectrum envelope (PSE) analyzing part 108. Moreover, this device is provided with a spectrum envelope reproducing part 111, and impulse response converting part 113, a DC component correcting part 115, a superposing part 117, a buffer memory 118 of two faces for an output, and a digital-analog (D/A) converter 119. In such a state, a low- pass component of voice spectrum is operated by the inpulse response converting part 113 of the voice spectrum. In such a way, a synthetic voice of a high quality can stably be obtd.

8. 发明专利

JPH02183300A SPEECH RECOGNITION DEVICE 失效
公开(公告)号：JPH02183300A
公开(公告)日：1990-07-17
申请号：JP338689
申请日：1989-01-10
申请人： HITACHI LTD , HITACHI SOFTWARE ENG
发明人： HATAOKA NOBUO , NAKASHIGE AKIRA , ARITSUKA TOSHIYUKI , AMANO AKIO , ICHIKAWA HIROSHI
IPC分类号： G06F15/18 , G06N3/00 , G06N99/00 , G10L15/02 , G10L15/08 , G10L15/16
摘要： PURPOSE:To speedily and accurately estimate the contents of a word and a phrase by inputting a phoneme or syllable string and correcting its recognition error, and estimating the contents of the word or phrase of an input speech. CONSTITUTION:This device is provided with a 1st-stage neural network which inputs the phoneme or syllable string obtained by a recognition processing means 3 and corrects the influence of acoustic coupling and a recognition error depending upon the recognition processing means and a 2nd-stage neural network 63 which estimates the contents of the word, phrase, etc., of the input speech. Then the recognition error is corrected and a phoneme string or syllable string which is corrected to a certain extent is inputted to estimate the contents of the word, phrase, etc. Consequently, the contents of the word or phrase of the input speech are accurately estimated fast from the phoneme or syllable string.

9. 发明专利

JPH0219960A NEURAL NETWORK AND METHOD AND DEVICE FOR ANALYSIS OF ACOUSTIC SIGNAL USING THE NETWORK 失效
公开(公告)号：JPH0219960A
公开(公告)日：1990-01-23
申请号：JP16870788
申请日：1988-07-08
申请人： HITACHI LTD
发明人： ICHIKAWA HIROSHI , AMANO AKIO , ARITSUKA TOSHIYUKI
IPC分类号： G06F15/18 , G06N3/00 , G06N99/00 , G10L11/00 , G10L15/02 , G10L15/08 , G10L15/16
摘要： PURPOSE:To facilitate analysis of the obtained features by forming each layer of a neural network of >=4 layers so as to analyze the layer structure for each factor and to realize the input for learning with a teacher to the intermediate layer from outside. CONSTITUTION:A 1st layer 24 of a neural network is formed into a 2-dimensional matrix corresponding to the basic physical feature values of the sound signal. The elements of a 2nd layer 25 are connected to all elements of the layer 24. Then the physical meaning of each element of the layer 25 is easily interpreted by checking the basic physical value of the elements of the layer 24 connected to each element of the layer 25 in a high degree of connection when the learning is through with a neuron net. In addition, some or all elements of a 3rd layer 26 are formed so as to reflect the voice element features in terms of the abstract meaning. The external learning is directly inputted in parallel with the learning value received from a 4th layer 27 by the back propagation. Thus the learning is performed in the form of the functions of said two types of learning. Thus it is possible to analyze the relation between the input voices and the output accordant with the purpose in response to the clear physical value.

10. 发明专利

JPH0214000A VOICE RECOGNIZING DEVICE 失效
公开(公告)号：JPH0214000A
公开(公告)日：1990-01-18
申请号：JP16250988
申请日：1988-07-01
申请人： HITACHI LTD
发明人： AMANO AKIO , HATAOKA NOBUO , ICHIKAWA HIROSHI
IPC分类号： G06F15/18 , G06N3/00 , G06N99/00 , G10L15/02 , G10L15/08 , G10L15/16
摘要： PURPOSE:To improve the voice recognition accuracy and to reduce the burden of a language processing part by providing a post-processing part for selecting, deleting, adding and correcting a recognition candidate. CONSTITUTION:The title device consists of a voice recognition processing part 1 for inputting an input voice 3 which a speaker has uttered and outputting a recognition result candidate sequence 4, and a hierarchical neural network 2 for inputting the recognition result candidate sequence 4 and outputting a correction recognition result 5. Also, by providing an output unit corresponding to an omission and an insertion on an output layer of the neural network 2, even when there are the omission and the insertion in the recognition result candidate sequence, the corresponding output pattern (teacher pattern) is shown on the output layer. Accordingly, in what kind of state the omission and the insertion are generated can be learned in the neural network 2. In such a way, a voice recognition result having high accuracy is obtained, and also, the burden of a language processing part can be reduced.

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