公开(公告)号：JP2001356795A

公开(公告)日：2001-12-26

申请号：JP2000176894

申请日：2000-06-13

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: G10L15/20 ,  G10L15/14 ,  G10L21/02

Abstract: PROBLEM TO BE SOLVED: To provide a voice recognition device and a voice recognizing method capable of performing voice recognition adapted to noise without preparing the model adapted to noise especially. SOLUTION: A feature vector component adjusting part 4 estimates the dispersion of a hidden Markov model by taking an amount of noise in the input voice into consideration based on the amount of noise which is included in an input voice and adjusts the featured values of the input voice in accordance with the estimated dispersion. A comparator part 14 recognizes the input voice based on hidden Markov models 151 to 15n in which the noise is not taken into consideration by using the adjusted featured values.

公开(公告)号：JPH0937258A

公开(公告)日：1997-02-07

申请号：JP18237695

申请日：1995-07-19

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N7/32

Abstract: PROBLEM TO BE SOLVED: To avoid reduction in the compression rate while suppressing a heavy load. SOLUTION: A motion detector 103 divides image data of a current frame outputted from an image input device 101 and image data of a preceding frame stored in a frame memory 110 into blocks with cross reference to each other. Then while moving a noted moving point on a preceding frame and its coordinate moving point, an area deformed by the moving is compensated and the compensated area is compared with an area of a corresponding current frame to obtain an error and a coordinate minimizing the error is set to each moving point of the preceding frame. In this case, when the entire error between frames is not within a permissible range, a moving is added newly to reduce the block size and an error minimum position of the moving point added newly is continuously detected.

公开(公告)号：JPH0937256A

公开(公告)日：1997-02-07

申请号：JP17997095

申请日：1995-07-17

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N19/50 ,  H04N19/105 ,  H04N19/137 ,  H04N19/176 ,  H04N19/186 ,  H04N19/196 ,  H04N19/20 ,  H04N19/42 ,  H04N19/423 ,  H04N19/46 ,  H04N19/463 ,  H04N19/503 ,  H04N19/51 ,  H04N19/517 ,  H04N19/533 ,  H04N19/537 ,  H04N19/57 ,  H04N19/61 ,  H04N19/625 ,  H04N19/70 ,  H04N19/80 ,  H04N19/91 ,  H04N7/32

Abstract: PROBLEM TO BE SOLVED: To reduce processing load while suppressing reduction in a compression rate when warp type motion compensation is predicted. SOLUTION: A motion detector 103 confirms position relation of moving points between frames by dividing image data of a current frame outputted from an image input deice 101 and image data of a preceding frame stored in a frame memory 110 into blocks, comparing them so as to detect an entire movement of an object on an image. A mask area is generated for each moving point depending on an edge detected on the current frame. Then a deformed mask area by movement is compensated while shifting a coordinate of the moving point noted on the preceding frame and the compensated mask area is compared with a mask area corresponding to the current frame to obtain an error between them and the coordinate minimizing the error is set as each moving point of the preceding frame.

公开(公告)号：JPH0670304A

公开(公告)日：1994-03-11

申请号：JP23893492

申请日：1992-08-14

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N7/30 ,  H04N7/32 ,  H04N7/133 ,  H04N7/137

Abstract: PURPOSE:To simplify the structure of the compressor by predicting motion based on a weight mean component calculated for frequency shift on a Wavelet conversion axis so as to eliminate the need for the processing of conversion time shift conversion on the conversion axis. CONSTITUTION:The compressor is provided with a Wavelet converter 51 applying Wavelet conversion to picture data and a weight mean device 54 taking weight mean of conversion plane data on a conversion axis loop for motion prediction, and the weight mean device 54 divides internally conversion axis data to calculate shift data and the Wavelet converter shifts the phase (shifts the frequency). Since the shift is calculated on the conversion axis, the processing of conversion time shift conversion having been required for a conventional compressor is eliminated. Then the prediction on the conversion axis loop is simply implemented regardless of the elimination of the processing of conversion, shift and inverse conversion on the conversion axis and then the configuration of the compressor is remarkably simplified.

公开(公告)号：JPH04223786A

公开(公告)日：1992-08-13

申请号：JP40684190

申请日：1990-12-26

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N19/60 ,  H04N19/102 ,  H04N19/136 ,  H04N19/423 ,  H04N19/48 ,  H04N19/625 ,  H04N19/91

Abstract: PURPOSE:To execute encoding with a pattern in which the bias of a code is large and to enhance encoding efficiency. CONSTITUTION:A picture compression device which block-divides and encodes picture information is provided with conversion parts 51-53 converting the arrangement of elements in a block and a selection part 54 selecting and encoding either the original block before conversion or the converted block by the conversion parts 51-53.

公开(公告)号：JPH04220878A

公开(公告)日：1992-08-11

申请号：JP40516090

申请日：1990-12-21

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N1/41 ,  G06T1/00 ,  G06T9/00 ,  H04N19/00 ,  H04N19/126 ,  H04N19/149 ,  H04N19/176 ,  H04N19/196 ,  H04N19/42 ,  H04N19/423 ,  H04N19/625 ,  H04N19/85

Abstract: PURPOSE:To decide a proper quantization map and to reduce deterioration in picture quality when a sample point is less by obtaining a function between a code quantity and a scale so as to decide the scale. CONSTITUTION:A picture data subject to discrete cosine transformation(DCT) by a DCT device 12 is quantized by using a quantization map to measure the quantity of a code quantity. A function of the code quantity and a scale is sampled exponentially and a function between the code quantity at a prediction point between sample points and the scale and a function of the code quantity by linear interpolation from the sample point and the scale are obtained to decide a scale SC. The picture data subject to DCT is quantized by using a new quantization map based on the scale SC decided in this way.

公开(公告)号：JPH04220876A

公开(公告)日：1992-08-11

申请号：JP40499490

申请日：1990-12-21

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N1/41 ,  G06T1/00 ,  G06T9/00 ,  H04N19/00 ,  H04N19/126 ,  H04N19/149 ,  H04N19/176 ,  H04N19/196 ,  H04N19/423 ,  H04N19/625 ,  H04N19/85

Abstract: PURPOSE:To increase the entire code quantity by setting each number of a quantization map to be a prescribed value or over so as to decrease a weight difference of each frequency thereby reducing a scale even when a bit width after quantization has a limit. CONSTITUTION:A data recorded in a RAM 22 of a basic quantization map is multiplied with a scale SC outputted from a scale calculation device in a 1st computing element 25. The result is inputted to a limiter 26 and when the result of multiplication is smaller than 128, 128 is outputted and when the result of multiplication is larger than 128, the result of multiplication is outputted. A picture data subject to discrete cosine transformation is quantized by using the output.

公开(公告)号：JP2002140058A

公开(公告)日：2002-05-17

申请号：JP2001222377

申请日：2001-07-24

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUOKA TAKESHI ,  MATSUI SHINICHI ,  IIZUKA NORIO ,  KASAHARA TAKESHI ,  MATSUDA TAKASHI ,  NAKAGOME KOICHI

IPC: H04N5/74 ,  G09G5/00

Abstract: PROBLEM TO BE SOLVED: To provide an image display device, etc., with excellent operability which can provide a proper display screen even when an image pickup device is placed at an arbitrary position and correctly reflect the position of an image picked-up spot (auxiliary light) in the display screen. SOLUTION: An image input part 2 which picks up an image of the display screen picks up an image of the display screen of a display part 3 and generates and supplies image data to a data processor 1. The data processor 1 extracts the position corresponding to a part in the display screen of the display part 3 that spot light emitted by a light emission part 4 impinges on and specifies which position in the display screen the position is through coordinate transformation. Then the display part 3 displays the spot at the specified position.

公开(公告)号：JPH11155147A

公开(公告)日：1999-06-08

申请号：JP26539598

申请日：1998-09-18

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N5/92 ,  H04N5/937 ,  H04N19/423 ,  H04N19/50 ,  H04N19/503 ,  H04N19/51 ,  H04N19/61 ,  H04N19/625 ,  H04N7/32

Abstract: PROBLEM TO BE SOLVED: To provide a image reproduction method in which the arithmetic operation of each block in a frame is devised to properly omit it, so as to reduce the arithmetic amount and coding and decoding of a bit stream executed at high speed by the communication media use coding algorithm, and to provided the image coding method and the image coder. SOLUTION: First, all blocks that are not referenced from any block in a Frame C are detected from among blocks of a Frame B (figure shows only a block b1 ), based on motion vectors mv(C0 )-mv(C3 ) resulting from decoding all motion vectors that reference the Frame B to be skipped, an IDCT arithmetic operation to the detected blocks is omitted so as to reduce the arithmetic amount. In an example shown in the figure, the IDCT arithmetic operation required for four blocks in the Frame B and four blocks in the Frame C substantially is reduced to 7 blocks by omitting the IDCT arithmetic operation with respect to the block b1 .

公开(公告)号：JPH10191347A

公开(公告)日：1998-07-21

申请号：JP35171696

申请日：1996-12-27

Applicant: CASIO COMPUTER CO LTD

Inventor： MATSUI SHINICHI

IPC: H04N19/50 ,  H04N19/423 ,  H04N19/46 ,  H04N19/463 ,  H04N19/503 ,  H04N19/51 ,  H04N19/513 ,  H04N19/517 ,  H04N19/533 ,  H04N19/57 ,  H04N19/61 ,  H04N19/625 ,  H04N19/70 ,  H04N19/91 ,  H04N7/32

Abstract: PROBLEM TO BE SOLVED: To reduce the arithmetic amount to detect a motion vector by detecting the rough motion of a semi-global block and detecting the motion of each block in the semi-global block to detect the motion vector of the block. SOLUTION: A controller 2 outputs an image mode signal, a prediction mode signal, a motion vector and various control signals to control the entire system. In the motion vector detection processing of the controller 2, two blocks are collected into a semi-global block and a rough motion is detected or predicted, the motion of each block in the semi-global block is detected to discriminate the motion of each block, then it is possible to reduce the number of times of arithmetic operations required to detect a motion vector. As a result, the motion vector is detected at a high speed and at a low cost because no large capacity memory for an arithmetic operation is required.