Claims Analysis
Total Claims: 14
Claim 1
Independent Claim
226 words
A vector quantization encoder comprising: a scan converter for converting an input image sequence into a block-formatted sequence; means for generating a data compression signal indicative of the amount of moving blocks in said block-formatted sequence; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differential image sequence applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; means connected to the outputs of said vector quantizers for generating a predictive image sequence; subtractor means for detecting a difference between said block-formatted sequence and said predictive image sequence and applying the detected difference to said vector quantizers as said interframe differential image sequence; and means connected to said vector quantizers for encoding said index signals.
Claim 2
Dependent Claim
226 words
A vector quantization encoder comprising: a scan converter for converting an input image sequence into a block-formatted sequence; means for generating a data compression signal indicative of the amount of moving blocks in said block-formatted sequence; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differential image sequence applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; means connected to the outputs of said vector quantizers for generating a predictive image sequence; subtractor means for detecting a difference between said block-formatted sequence and said predictive image sequence and applying the detected difference to said vector quantizers as said interframe differential image sequence; and means connected to said vector quantizers for encoding said index signals.
Claim 3
Dependent Claim
149 words
A vector quantization encoder as claimed in claim 1, wherein said output vectors of each of said vector quantizers are prepared in accordance with the steps of: (a) performing orthogonal transform on interframe differential training image sequences, so that said orthogonally transformed interframe differential training image sequences are mapped onto a frequency domain; (b) partitioning the mapped training image sequences into a plurality of frequency bands; (c) deriving a code table of optimum quantized vectors from the mapped orthogonally transformed interframe differential training image sequences which are partitioned into the frequency band corresponding to said particular frequency band; and (d) performing inverse orthogonal transform on said optimum quantized vectors, so that the inversely orthogonally transformed optimum quantized vectors are mapped onto a spatial domain as said output vectors, said spatial domain defining storage locations of said memory which are addressable as a function of said interframe differential image sequence.
Claim 4
Dependent Claim
302 words
A vector quantization encoder comprising: a scan converter for converting an input image sequence into a block-formatted sequence; a motion detector for detecting moving blocks in said block-formatted sequence; an integrator for integrating said detected moving blocks; comparator means for comparing the output of said integrator with one or more reference values and deriving therefrom a data compression signal indicative of the amount of said detected moving blocks; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differerential image sequence, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; a plurality of inverse vector quantizers for performing inverse vector quantization on the index signals from said vector quantizers; adder means for summing said index signals; a frame memory for introducing a frame interval delay to an output signal from said adder means and supplying the delayed signal to said adder means to produce a prediction image sequence; subtractor means for detecting a difference between the outputs of said scan converter and said frame memory for supplying the detected difference to said vector quantizers as said interframe differential image sequence; and encoder means connected to said vector quantizers for encoding said index signals.
Claim 5
Dependent Claim
302 words
A vector quantization encoder comprising: a scan converter for converting an input image sequence into a block-formatted sequence; a motion detector for detecting moving blocks in said block-formatted sequence; an integrator for integrating said detected moving blocks; comparator means for comparing the output of said integrator with one or more reference values and deriving therefrom a data compression signal indicative of the amount of said detected moving blocks; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differerential image sequence, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; a plurality of inverse vector quantizers for performing inverse vector quantization on the index signals from said vector quantizers; adder means for summing said index signals; a frame memory for introducing a frame interval delay to an output signal from said adder means and supplying the delayed signal to said adder means to produce a prediction image sequence; subtractor means for detecting a difference between the outputs of said scan converter and said frame memory for supplying the detected difference to said vector quantizers as said interframe differential image sequence; and encoder means connected to said vector quantizers for encoding said index signals.
Claim 6
Dependent Claim
149 words
A vector quantization encoder as claimed in claim 3, wherein said output vectors of each of said vector quantizers are prepared in accordance with the steps of: (a) performing orthogonal transform on interframe differential training image sequences, so that said orthogonally transformed interframe differential training image sequences are mapped onto a frequency domain; (b) partitioning the mapped training image sequences into a plurality of frequency bands; (c) deriving a code table of optimum quantized vectors from the mapped orthogonally transformed interframe differential training image sequences which are partitioned into the frequency band corresponding to said particular frequency band; and (d) performing inverse orthogonal transform on said optimum quantized vectors, so that the inversely orthogonally transformed optimum quantized vectors are mapped onto a spatial domain as said output vectors, said spatial domain defining storage locations of said memory which are addressable as a function of said interframe differential image sequence.
Claim 7
Dependent Claim
75 words
A vector quantization encoder as claimed in claim 3, wherein said motion detector detects motion vector and controls the timing of said encoder means in accordance with the detected motion vector, and said decoder means generates a timing signal representative of the detected motion vector, further comprising: variable delay means connected between said subtractor and said frame memory for introducing a variable delay to said prediction image sequence in accordance with the detected motion vector.
Claim 8
Dependent Claim
280 words
A communication system comprising: a scan converter for converting an input image sequence into a block-formatted sequence; means for generating a data compression signal indicative of the amount of moving blocks in said block-formatted sequence; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differential image sequence applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; means connected to the outputs of said vector quantizers for generating a predictive image sequence; subtractor means for detecting a difference between said block-formatted sequence and said predictive image sequence and applying the detected difference to said vector quantizers as said interframe differential image sequence; encoder means connected to said vector quantizers for encoding said index signals and transmitting the encoded signals through a transmission medium; decoder means connected to said transmission medium for decoding said transmitted signals; a plurality of inverse vector quantizers for performing inverse vector quantization on said decoded signals; and means connected to the outputs of said inverse vector quantizers for generating a replica of said block-formatted sequence.
Claim 9
Dependent Claim
280 words
A communication system comprising: a scan converter for converting an input image sequence into a block-formatted sequence; means for generating a data compression signal indicative of the amount of moving blocks in said block-formatted sequence; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of an interframe differential image sequence applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; means connected to the outputs of said vector quantizers for generating a predictive image sequence; subtractor means for detecting a difference between said block-formatted sequence and said predictive image sequence and applying the detected difference to said vector quantizers as said interframe differential image sequence; encoder means connected to said vector quantizers for encoding said index signals and transmitting the encoded signals through a transmission medium; decoder means connected to said transmission medium for decoding said transmitted signals; a plurality of inverse vector quantizers for performing inverse vector quantization on said decoded signals; and means connected to the outputs of said inverse vector quantizers for generating a replica of said block-formatted sequence.
Claim 10
Dependent Claim
390 words
A communication system comprising: a scan converter for converting an input image sequence into a block-formatted sequence; a motion detector for detecting moving blocks in said block-formatted sequence; an integrator for integrating said detected moving blocks; comparator means for comparing the output of said integrator with one or more reference values and deriving therefrom a data compression signal indicative of the amount of said detected moving blocks; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of a prediction error applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; a plurality of inverse vector quantizers for performing inverse vector quantization on the index signals from said vector quantizers; first adder means for summing said index signals; a first frame memory for introducing a frame interval delay to an output signal from said adder means and supplying the delayed signal to said first adder means to produce a predictive image sequence; subtractor means for detecting a difference between the outputs of said scan converter and said frame memory for supplying the detected difference to said vector quantizers as said interframe differential image sequence; encoder means connected to said vector quantizers for encoding said index signals and transmitting the encoded signals through a transmission medium; decoder means connected to said transmission medium for decoding said transmitted signals; a plurality of inverse vector quantizers for performing inverse vector quantization on said decoded signals; second adder means for summing the outputs of said inverse vector quantizers; and a second frame memory for introducing a frame interval delay to an output signal from said second adder means and applying the delayed output signal to said second adder means to generate a replica of said block-formatted sequence.
Claim 11
Dependent Claim
390 words
A communication system comprising: a scan converter for converting an input image sequence into a block-formatted sequence; a motion detector for detecting moving blocks in said block-formatted sequence; an integrator for integrating said detected moving blocks; comparator means for comparing the output of said integrator with one or more reference values and deriving therefrom a data compression signal indicative of the amount of said detected moving blocks; a plurality of vector quantizers individually selectively operable in accordance with said data compression signal, each of said vector quantizers having a particular frequency band and a memory containing output vectors, said output vectors of each of said vector quantizers being representative of inverse orthogonal transform of a code table of optimum quantized vectors in said particular frequency band, said optimum quantized vectors being orthogonal transform of interframe differential training image sequences, said output vectors being readable from said memory as a function of a prediction error applied thereto, each of said vector quantizers further including means for selecting one of the output vectors read out of said memory which is nearest to the value of said interframe differential image sequence and generating an index signal representative of the selected vector; a plurality of inverse vector quantizers for performing inverse vector quantization on the index signals from said vector quantizers; first adder means for summing said index signals; a first frame memory for introducing a frame interval delay to an output signal from said adder means and supplying the delayed signal to said first adder means to produce a predictive image sequence; subtractor means for detecting a difference between the outputs of said scan converter and said frame memory for supplying the detected difference to said vector quantizers as said interframe differential image sequence; encoder means connected to said vector quantizers for encoding said index signals and transmitting the encoded signals through a transmission medium; decoder means connected to said transmission medium for decoding said transmitted signals; a plurality of inverse vector quantizers for performing inverse vector quantization on said decoded signals; second adder means for summing the outputs of said inverse vector quantizers; and a second frame memory for introducing a frame interval delay to an output signal from said second adder means and applying the delayed output signal to said second adder means to generate a replica of said block-formatted sequence.
Claim 12
Dependent Claim
114 words
A communication system as claimed in claim 7, wherein said motion detector detects motion vector and controls the timing of said encoder means in accordance with the detected motion vector, and said decoder means generates a timing signal representative of the detected motion vector, further comprising: first variable delay means connected between said subtractor and said frame memory for introducing a variable delay to said predictive image sequence in accordance with said detected motion vector; and second variable delay means connected between said second adder means and said second frame memory for introducing a variable delay to the output of said second frame memory in accordance with said timing signal from said decoder means.
Claim 13
Independent Claim
101 words
A method for producing code tables, comprising the steps of: (a) performing orthogonal transform on interframe differential training image sequences, so that said orthogonally transformed interframe training image sequences are mapped onto a frequency domain; (b) partitioning the mapped training image sequences into different frequency bands; (c) deriving a plurality of code tables of optimum quantized vectors respectively from the partitioned training image sequences; and (d) performing inverse orthogonal transform on said quantized vectors of each of said code tables, so that the inversely orthogonally transformed optimum quantized vectors of each of said code tables are mapped onto a spatial domain.
Claim 14
Independent Claim
101 words
A method for producing code tables, comprising the steps of: (a) performing orthogonal transform on interframe differential training image sequences, so that said orthogonally transformed interframe training image sequences are mapped onto a frequency domain; (b) partitioning the mapped training image sequences into different frequency bands; (c) deriving a plurality of code tables of optimum quantized vectors respectively from the partitioned training image sequences; and (d) performing inverse orthogonal transform on said quantized vectors of each of said code tables, so that the inversely orthogonally transformed optimum quantized vectors of each of said code tables are mapped onto a spatial domain.