A new method for video data compression by quadratic Bézier curve fitting

This paper presents a new and efficient method for video data compression using quadratic Bézier curve fitting. The method treats the luminance or color variations of a spatial location in a sequence of frames as input points in Euclidean space R1 or R3. The input points are approximated using quadratic Bézier least square fitting. The output data consists of quadratic Bézier control points and difference between original and fitted data. Video data compression is the main application of proposed method. It is shown that entropy of output data is significantly less than the entropy of input data. The method can be applied to 1-D space like luminance and chrominance components separately or 3-D color spaces such as RG B and Y CbCr .