RECOGNITION OF FACE USING FEATURE VECTOR WITH DCT AS BASE

    公开(公告)号:JPH1055444A

    公开(公告)日:1998-02-24

    申请号:JP11546197

    申请日:1997-05-06

    Abstract: PROBLEM TO BE SOLVED: To enable the recognition and identification of a human face especially included in an image by collating a prescribed object with one of known objects based on the error of coding. SOLUTION: A preprocessor 21 extracts a concerned area (namely face) for an inputted image and standardizes the scale of reduction. The extracted section in the inputted image is divided into blocks and the respective blocks are converted with a clock as a base by transformation 22. When a DCT vector is once provided by the transformation 22, a sorter 24 is used for identifying the face included in the inputted face based on code books from 23-1 or 23-3 of the known face. For example, the sorter 24 judges which code books is most coincident with an input vector provided from the inputted image. In the case of face identification, the face related to the 'most coincident' code book is identified as the strongest candidate.

    METHOD AND SYSTEM FOR PROVIDING CHANNEL ERROR PROTECTION TO BIT STREAM

    公开(公告)号:JP2001203588A

    公开(公告)日:2001-07-27

    申请号:JP2000349289

    申请日:2000-11-16

    Abstract: PROBLEM TO BE SOLVED: To avoid a problem of excess overhead due to a tail bit required for termination of a sub block in a conventional ununiform error protection(UEP) technology while avoiding error propagation in a channel decoder. SOLUTION: The UEP is provided to a source coded bit stream 200 by a puncturing of a channel code. The channel code has a designated rate relating to a corresponding error protection level and is used for the purpose of coding a 1st part 202 of the source coded bit stream 200. Then the channel code is punctured for at least a 2nd part 204 of the source coded bit stream 200 so as to obtain a rate higher than the designated rate. The start of the puncturing is intentionally delayed by an amount decided on the basis of at least partially the memory of the channel code.

    METHOD AND APPARATUS FOR CODING CHROMINANCE INFORMATION IN COLOR VIDEO ENCODER

    公开(公告)号:JPH09163405A

    公开(公告)日:1997-06-20

    申请号:JP22351396

    申请日:1996-08-26

    Abstract: PROBLEM TO BE SOLVED: To obtain a coding technology for a color video signal used for video communication employing a very low bit transmission speed by bypassing entry of a code book identified by indexing the code book for coding chrominace information when the bypassing is discriminated to be advantageous. SOLUTION: Converted luminance and chrominance blocks are subject to vector quantization by a quantizer 15 by using a code book 25. Then a code book index signal is generated based on luminance information of the coded block and a coding error is decided based on the comparison between chrominance information included in the given block and a chrominance vector displayed by a generated code book index signal. A chrominance spectrum bypass signal is generated based on whether or not the coding error exceeds the threshold level.

    4.
    发明专利
    未知

    公开(公告)号:DE60032714T2

    公开(公告)日:2007-11-08

    申请号:DE60032714

    申请日:2000-11-06

    Abstract: Unequal channel error protection is provided for a bit stream, e.g., a progressive source coded bit stream, by puncturing of a channel code. The channel code has a specified rate associated with a corresponding level of error protection, and is used to code a first portion of the bit stream. The channel code is subsequently punctured for at least a second portion of the bit stream so as to provide a rate higher than that of the designated rate. Commencement of the puncturing is deliberately delayed by an amount determined based at least in part on a memory of the channel code. Advantageously, the delay prevents propagation of errors in a channel decoder used to decode the channel coded bit stream. The progressive bit stream may be a progressive source coded video or audio bit stream, any other type of progressive information bit stream, or a partitioned source coded bit stream.

    5.
    发明专利
    未知

    公开(公告)号:DE69718553T2

    公开(公告)日:2003-08-14

    申请号:DE69718553

    申请日:1997-04-30

    Abstract: A method and apparatus for recognition of objects such as faces in images using signal compression techniques (i.e., coding techniques) in which a portion of the image which includes the object to be recognized (e.g., the face) is coded, and the resultant coded data is matched against previously coded and stored training data which makes up a known object database. A given object in an input image signal is matched to one of a plurality of known objects stored in a database, wherein the stored representation of each of the known objects comprises a codebook generated based on training image signals comprising the known object. A first illustrative embodiment comprises the steps of decomposing the given object into blocks; performing a plurality of encodings of the given object, each encoding comprising coding the object with use of one of the codebooks; determining a coding error for each encoding; and matching the given object to one of the known object based on the coding errors. A second illustrative embodiment comprises the steps of decomposing the given object into blocks; generating a codebook corresponding to the given object based on the blocks; comparing the codebook corresponding to the given object with the codebooks corresponding to each of the known objects; and matching the given object to one of the known objects based on the comparison of the codebooks.

    METHOD AND APPARATUS FOR PROVIDING CHANNEL ERROR PROTECTION FOR A SOURCE CODED BIT STREAM

    公开(公告)号:CA2325431A1

    公开(公告)日:2001-05-16

    申请号:CA2325431

    申请日:2000-11-07

    Abstract: Unequal channel error protection is provided for a bit stream, e.g., a progressive source coded bit stream, by puncturing of a channel code. The channel code h as a specified rate associated with a corresponding level of error protection, and is used to code a first portion of the bit stream. The channel code is subsequently punctured for at least a second portion of the bit stream so as to provide a rate higher than that of the designated rate. Commencement of the puncturing is deliberately delayed by an amount determin ed based at least in part on a memory of the channel code. Advantageously, the delay prevents propagation of errors in a channel decoder used to decode the channel coded bit stream. The progressive bit stream may be a progressive source coded video or audio bit stream, any other type of progressive information bit stream, or a partitioned source coded bi t stream.

    7.
    发明专利
    未知

    公开(公告)号:DE60032714D1

    公开(公告)日:2007-02-15

    申请号:DE60032714

    申请日:2000-11-06

    Abstract: Unequal channel error protection is provided for a bit stream, e.g., a progressive source coded bit stream, by puncturing of a channel code. The channel code has a specified rate associated with a corresponding level of error protection, and is used to code a first portion of the bit stream. The channel code is subsequently punctured for at least a second portion of the bit stream so as to provide a rate higher than that of the designated rate. Commencement of the puncturing is deliberately delayed by an amount determined based at least in part on a memory of the channel code. Advantageously, the delay prevents propagation of errors in a channel decoder used to decode the channel coded bit stream. The progressive bit stream may be a progressive source coded video or audio bit stream, any other type of progressive information bit stream, or a partitioned source coded bit stream.

    METHOD AND APPARATUS FOR PROVIDING CHANNEL ERROR PROTECTION FOR A SOURCE CODED BIT STREAM

    公开(公告)号:CA2325431C

    公开(公告)日:2005-09-20

    申请号:CA2325431

    申请日:2000-11-07

    Abstract: Unequal channel error protection is provided for a bit stream, e.g., a progressive source coded bit stream, by puncturing of a channel code. The channel code h as a specified rate associated with a corresponding level of error protection, and is used to code a first portion of the bit stream. The channel code is subsequently punctured for at least a second portion of the bit stream so as to provide a rate higher than that of the designated rate. Commencement of the puncturing is deliberately delayed by an amount determin ed based at least in part on a memory of the channel code. Advantageously, the delay prevents propagation of errors in a channel decoder used to decode the channel coded bit stream. The progressive bit stream may be a progressive source coded video or audio bit stream, any other type of progressive information bit stream, or a partitioned source coded bi t stream.

    9.
    发明专利
    未知

    公开(公告)号:DE69718553D1

    公开(公告)日:2003-02-27

    申请号:DE69718553

    申请日:1997-04-30

    Abstract: A method and apparatus for recognition of objects such as faces in images using signal compression techniques (i.e., coding techniques) in which a portion of the image which includes the object to be recognized (e.g., the face) is coded, and the resultant coded data is matched against previously coded and stored training data which makes up a known object database. A given object in an input image signal is matched to one of a plurality of known objects stored in a database, wherein the stored representation of each of the known objects comprises a codebook generated based on training image signals comprising the known object. A first illustrative embodiment comprises the steps of decomposing the given object into blocks; performing a plurality of encodings of the given object, each encoding comprising coding the object with use of one of the codebooks; determining a coding error for each encoding; and matching the given object to one of the known object based on the coding errors. A second illustrative embodiment comprises the steps of decomposing the given object into blocks; generating a codebook corresponding to the given object based on the blocks; comparing the codebook corresponding to the given object with the codebooks corresponding to each of the known objects; and matching the given object to one of the known objects based on the comparison of the codebooks.

Patent Agency Ranking