Three-dimensional Epanechnikov mixture regression in image coding

Kernel methods have been studied extensively in recent years. We propose a three-dimensional (3-D) Epanechnikov Mixture Regression (EMR) based on our (EK) and realize complete framework for image coding. In research, we deduce the covariance-matrix form of 3-D kernels their correlated statistics to obtain mixture models. To apply theories coding, EMR which can better model an smaller blocks compared with conventional Gaussian (GMR). The regressions are all improved Expectation-Maximization (EM) algorithm mean square error optimization. Finally, design Adaptive Mode Selection (AMS) best pattern combination Our recovered has clear outlines superior coding efficiency JPEG below 0.25bpp. work realizes unprecedented theory application by: (1) enriching kernel,(2) improving EM using MSE optimization, (3) exploiting its (4) AMS optimal modeling combined kernel.