Multidimensional video processing

摘要

A method and system for multidimensional video processing comprises preprocessing a video signal in at least two of horizontal, vertical and temporal dimensions of the video signal to reduce an amount of information of the video signal to obtain an information-reduced signal and filtered components that are separated from the video signal; and using the information-reduced signal to generate simulated filtered components, and generating support signals based on a difference between the filtered components and the simulated filtered components.

主权项

1. A method for multidimensional video processing, the method comprising:
preprocessing a video signal in at least one of a horizontal dimension and a vertical dimension, and preprocessing the video signal in a temporal dimensions to reduce an amount of information of the video signal to obtain an information-reduced signal and filtered components that are separated from the video signal; and using the information-reduced signal to generate simulated filtered components, and generating support signals based on a difference between the filtered components and the simulated filtered components; wherein the preprocessing in the horizontal dimension comprises:
separating horizontal high and low frequencies from the video signal by horizontally low-pass filtering the video signal,outputting the information-reduced signal as horizontal-low frequencies, and outputting a horizontal filtered component comprising horizontal-high frequencies,generating simulated high frequencies from the horizontal-low frequencies, andsubtracting the simulated high frequencies subtracted from the horizontal high frequencies to generate a horizontal support signal; wherein the preprocessing in the vertical dimension comprises:
separating even and odd lines of the video signal and eliminating one line out of two,outputting the information-reduced signal comprising the even lines, and outputting a vertical filtered component comprising the odd lines;generating simulated odd lines from the even lines, andsubtracting the simulated odd lines from the odd lines to generate a vertical support signal; wherein the preprocessing in the temporal dimension comprises:
separating even and odd frames of the video signal to process the video signal in a time domain,outputting the information-reduced signal comprising the even frames, and outputting a temporal filtered component comprising the odd frames;generating simulated odd frames from the even frames, andoutputting a temporal support signal generated by subtracting the simulated odd frames subtracted from the odd frames;and outputting the information-reduced signal, the temporal support signal and at least one of the horizontal support signal and the vertical support signal such that less overall information is output than the video signal.