Total Variation Based Image Cartoon-Texture Decomposition

摘要: Abstract : This paper studies algorithms for decomposing a real image into the sum of cartoon and texture based on total variation minimization second-order cone programming (SOCP). The is represented as function bounded while (and noise) by elements in space oscillating functions, proposed Yves Meyer. Our approach gives more accurate results than those obtained previously Vese-Osher's approximation to Meyer's model, which we also formulate solve an SOCP. model minimizing with L1-norm fidelity term considered empirically shown achieve even better when there no noise. analyzed be able select features according their scales.