Apparatus for detecting motion vector and encoding picture signals

Claim 1

A method for comparing data representing a first sequential picture frame and a second sequential picture frame to detect a motion vector comprising the steps of: generating a first motion vector and first minimum value by block matching said data of said first sequential picture frame and said second sequential picture frame, wherein a luminance pattern is obtained by eliminating dc components from a luminance signal so as to yield only ac components; generating a second motion vector and second minimum value by block matching said data of said first sequential picture frame and said second sequential picture frame wherein the difference between said first sequential picture frame data and said second sequential picture frame data is calculated; evaluating said first motion vector and first minimum value, and said second motion vector and second minimum value to determine which meets predetermined criteria; outputting either said first motion vector and first minimum value or said second motion vector and second minimum value as an optimum motion vector and optimum minimum value, based upon said evaluation; and encoding video frame data based upon said optimum motion vector and minimum value.

Claim 2

A method for encoding a picture signal comprising the steps of: generating a first motion vector and minimum value generated by block matching employing a luminance pattern obtained by eliminating dc components from a luminance signal of a picture and thus containing only ac components, to a second motion vector generated by block matching employing a luminance signal of a picture; generating an optimum motion vector, based upon the result of comparison between said first motion vector and said second motion vector, as a true motion vector; effecting encoding and quantization on a prediction picture signal afforded by said true motion vector; and variable-length encoding the resulting quantized signal and the motion vector.

Claim 3

An apparatus for encoding a video signal comprising: input for receiving an input video signal on a sequential frame by frame basis; first memory means for storing first frame data representing a plurality of pixel luminance values; second memory means for storing second frame data representing a plurality of pixel luminance values; first arithmetic logic unit for calculating a difference between luminance values of said second frame data and said first frame data in accordance with the expression: Out₁ = |A - B| , where A represents a luminance value of said first frame data and B represents a luminance value of said second frame data, and outputting a signal representing same; circuitry for generating and outputting a first motion vector signal representing a first motion vector and minimum value in accordance with said output from said first arithmetic logic unit; first average luminance calculating unit for calculating the average luminance value of said first frame data; second average luminance calculating unit for calculating the average luminance value of said second frame data; second arithmetic logic unit for generating a second output signal in accordance with the expression: Out₂ = |(A - A avg. ) - (B - B avg. )| , where A avg. represents the total average luminance value of said first frame data and B avg. represents the total average luminance value of said second frame data; evaluation circuitry for comparing the value of said first output signal with the value of said second output signal to determine which value meets predetermined criteria; selection circuitry for outputting as an optimum motion vector and value the one of either said first output signal or said second output signal which is determined to meet said predetrmined criteria; and encoder circuitry for encoding said video signal in accordance with said optimum motion vector and value.

Claim 4

A method for detecting a motion vector comprising the steps of: storing first picture data into a first memory; storing second picture data into a second memory; said first picture data and said second picture data represent sequential frames of video images; generating a first motion vector based upon the difference in luminance values between said first picture data and said second picture data and wherein said dc components of said luminance values have been removed; generating a second motion vector based upon the difference of the difference in luminance values between said first picture data and said second picture data; comparing said first motion vector and said second motion vector to determine which has the lesser value; and outputting as an optimum motion vector, the one of either said first motion vector or said second motion vector having the lesser value.

Claim 5

An apparatus for detecting a motion vector between a present frame and a preceding frame of an input digital picture signal which includes DC components and AC components comprising: eliminating means for eliminating DC components from the input digital picture signal so as to provide a digital picture signal containing only AC components; first motion vector generating means for generating a first motion vector and first minimum value from the digital picture signal supplied from said eliminating means by utilizing block matching algorithm; second motion vector generating means for generating second motion vector and second minimum value from the input digital picture signal by utililzing block matching algorithm; evaluation means for evaluating said first and second motion vectors to determine which meets predetermined criteria; and selection means for selectively outputting said first motion vector or said second motion vector on the basis of said determination of said evaluation means.