Time-varying image encoder

Claim 2

The motion picture coding apparatus according to claim 1, wherein said second predicting means supplies a signal obtained by shifting said signal up-sampled by said up-sampling means by an amount corresponding to motion within a field period, to candidates for said low-resolution predictive signal.

Claim 3

A motion picture coding apparatus comprising: first coding means for performing predictive coding on a high-resolution picture; first local decoding means for decoding a coding result from said first coding means to obtain a high-resolution local decoded signal; first predicting means for attaining a high-resolution predictive signal from said high-resolution local decoded signal; converting means for converting said high-resolution picture to a low-resolution picture; second coding means for coding said low-resolution picture signal; second local decoding means for decoding a coding result from said second coding means to obtain a low-resolution local decoded signal; up-sampling means for up-sampling said low-resolution local decoded signal; and second predicting means for attaining a low-resolution predictive signal from a signal up-sampled by said up-sampling means, whereby as a prediction mode used when prediction-coding said high-resolution picture, a predictive signal is attained by a weighted sum of said low-resolution predictive signal and said high-resolution predictive signal.

Claim 4

The motion picture coding apparatus according to claim 3, wherein an amount of shift corresponding to motion in said field period is obtained by dividing a shift amount necessary to prepare said high-resolution predictive signal by an inter-field distance, and a signal produced from a low-resolution predictive signal obtained by said amount of shift and said high-resolution predictive signal is added to candidates for said predictive signal.

Claim 5

The motion picture coding apparatus according to claim 3, wherein an amount of shift corresponding to motion in said field period is obtained as an optimal shift amount to optimize an evaluation function attained by searching around a reference shift amount, obtained by dividing a shift amount necessary to prepare said high-resolution predictive signal by an inter-field distance, and a difference between said optimal shift amount and said reference shift amount is transmitted.

Claim 6

The motion picture coding apparatus according to claim 3 or 5, wherein said first predicting means has means for outputting, as said high-resolution predictive signal, a signal obtained by adding signals, undergone conversion having a mutually complementary characteristic with respect to both of said high-resolution local decoded signal and said low-resolution local decoded signal.

Claim 7

A motion picture coding apparatus comprising: first coding means for performing predictive coding on a high-resolution picture signal; first local decoding means for decoding a coding result from said first coding means to obtain a high-resolution local decoded signal; first predicting means for attaining a high-resolution predictive signal from said high-resolution local decoded signal; converting means for converting said high-resolution picture signal to a low-resolution picture signal; second coding means for coding said low-resolution picture signal; second local decoding means for decoding a coding result from said second coding means to obtain a low-resolution local decoded signal; up-sampling means for up-sampling said low-resolution local decoded signal; second predicting means for attaining a low-resolution predictive signal from a signal up-sampled by said up-sampling means; and prediction-signal selecting means for selecting a predictive signal used by said first coding means from said high-resolution predictive signal and said low-resolution predictive signal under a determination criterion where prediction error power in cases when said high-resolution predictive signal and said low-resolution predictive signal are used in predictive coding in said first coding means, and an amount of additional information generated by said first coding means are considered.

Claim 8

A motion picture coding apparatus comprising: first coding means having a mode for performing predictive coding on a high-resolution prediction error signal, which is a difference between a high-resolution picture signal and a high-resolution predictive signal, using a low-resolution prediction error signal; first local decoding means for decoding a coding result from said first coding means to obtain a high-resolution local decoded signal; first predicting means for detecting a motion vector between said high-resolution local decoded signal and said high-resolution picture signal, and attaining said high-resolution predictive signal undergone motion compensation using said motion vector; converting means for converting said high-resolution picture signal to a low-resolution picture signal; second coding means for coding said low-resolution picture signal; second local decoding means for decoding a coding result from said second coding means to obtain a low-resolution local decoded signal; up-sampling means for up-sampling said low-resolution local decoded signal; and second predicting means for attaining a low-resolution predictive signal, undergone motion compensation using said motion vector obtained by said first predicting means, from a signal up-sampled by said up-sampling means, and obtaining a difference between said low-resolution predictive signal and said low-resolution local decoded signal to yield said low-resolution prediction error signal.

Claim 9

A motion picture coding apparatus for performing predictive coding on a high-resolution picture signal and performing predictive coding of a low-resolution picture signal obtained by converting said high-resolution picture signal, comprising: first predicting means for producing a high-resolution predictive signal based on a high-resolution local decoded signal decoded from a signal resulting from predictive coding of said high-resolution picture signal; up-sampling means for up-sampling a low-resolution local decoded signal decoded from a signal resulting from predictive coding of said low-resolution picture; second predicting means for producing a low-resolution predictive signal based on an output signal of said up-sampling means; and third predicting means for producing one of said high-resolution predictive signal, said low-resolution predictive signal and signals obtained by weighting and adding said high-resolution predictive signal and said low-resolution predictive signal, as a predictive signal used when performing predictive coding on said high-resolution picture signal, said up-sampling means including, (a) intra-field interpolation-signal producing means for producing an intra-field interpolation signal consisting only of one field of a signal of said low-resolution local decoded signal, (b) intra-frame interpolation-signal producing means for producing an intra-frame interpolation signal consisting of consecutive two fields of signals of said low-resolution local decoded signal, and (c) selecting means for selecting, as said output signal of said up-sampling means, that one signal among said intra-field interpolation signal, said intra-frame interpolation signal and signals obtained by weighting and adding said interpolation signals, which is adaptively determined based on at least one of said high-resolution picture signal and other information necessary for coding.

Claim 10

A motion picture coding apparatus for performing predictive coding on a high-resolution picture signal and performing predictive coding of a low-resolution picture signal obtained by converting said high-resolution picture signal, comprising: first predicting means for producing a high-resolution predictive signal based on a high-resolution local decoded signal decoded from a signal resulting from predictive coding of said high-resolution picture signal; up-sampling means for up-sampling a low-resolution local decoded signal decoded from a signal resulting from predictive coding of said low-resolution picture; second predicting means for producing a low-resolution predictive signal based on an output signal of said up-sampling means; and third predicting means for producing one of said high-resolution predictive signal, said low-resolution predictive signal and signals obtained by weighting and adding said high-resolution predictive signal and said low-resolution predictive signal, as a predictive signal used when performing predictive coding on said high-resolution picture signal, said up-sampling means including, (a) first and second motion-compensation up-sampling means for performing motion-compensation up-sampling on consecutive two fields of said low-resolution local decoded signal, and (b) motion-vector detecting means for detecting a motion vector used in said first and second motion-compensation up-sampling means, referring to said high-resolution picture signal.

Claim 11

A motion picture coding apparatus for performing predictive coding on a high-resolution picture signal and performing predictive coding of a low-resolution picture signal obtained by converting said high-resolution picture signal, comprising: first predicting means for producing a high-resolution predictive signal based on a high-resolution local decoded signal decoded from a signal resulting from predictive coding of said high-resolution picture signal; up-sampling means for up-sampling a low-resolution local decoded signal decoded from a signal resulting from predictive coding of said low-resolution picture; second predicting means for producing a low-resolution predictive signal based on an output signal of said up-sampling means; and third predicting means for producing one of said high-resolution predictive signal, said low-resolution predictive signal and signals obtained by weighting and adding said high-resolution predictive signal and said low-resolution predictive signal, as a predictive signal used when performing predictive coding on said high-resolution picture signal, said up-sampling means including, (a) first and second motion-compensation up-sampling means for performing motion-compensation up-sampling on consecutive two fields of said low-resolution local decoded signal, and (b) vector computing means for computing a motion vector used in said first and second motion-compensation up-sampling means, based on at least one of a motion vector used in predictive coding of said low-resolution picture signal and a motion vector used in predictive coding of said high-resolution picture signal.

Claim 12

A motion picture coding apparatus for performing predictive coding on a high-resolution picture signal and performing predictive coding of a low-resolution picture signal obtained by converting said high-resolution picture signal, comprising: first predicting means for producing a high-resolution predictive signal based on a high-resolution local decoded signal decoded from a signal resulting from predictive coding of said high-resolution picture signal; up-sampling means for up-sampling a low-resolution local decoded signal decoded from a signal resulting from predictive coding of said low-resolution picture; second predicting means for producing a low-resolution predictive signal based on an output signal of said up-sampling means; third predicting means for producing one of said high-resolution predictive signal, said low-resolution predictive signal and signals obtained by weighting and adding said high-resolution predictive signal and said low-resolution predictive signal, as a predictive signal used when performing predictive coding on said high-resolution picture signal; and weighting-coefficient determining means for determining a weighting coefficient used in said weighting and adding in said third predicting means, using at least one of (a) a motion vector used in predictive coding of said low-resolution picture signal, (b) a motion vector used in predictive coding of said high-resolution picture signal, (c) a motion vector used in motion-compensation up-sampling when said up-sampling means performs said motion-compensation up-sampling, and (d) an amount of shift in a case of predicting a signal obtained by shifting a picture resulting from up-sampling of a low-resolution picture by an amount corresponding to motion within a field period.