Avoiding oscillation phenomenon which occurs in analysis and reconstruction of image edges, an application without extrapolation both sides is proposed. Image energy is concentrated in low frequency by using nonlinear transform, and the exact jumps location are detected by local maximum of wavelet transform in fine scale. Non-oscillation edges of reconstructed image are received with less storage space and computing time.
Starck J L, Murtagh, F, Bijaoui A. Image processing and data analysis[J]. Astronomy and Astrophysics Supplement Series, 2000:139- 149.[2]Donoho D. Wedgelets: Nearly-minimax estimation of edges[J].Annals of Stat.1999, 27(3):859-[3]Candes E, Donoho D. New tight frames of curvelets and optimal representations of objects with C2 Singularities[R]. Technical Report, Stanford University, 2002.[4]Chan T F. Adaptive ENO-wavelet transforms for discontinuous functions[R]. in 12th Domain Decomposition Conf, Paris, June 26 - 28, 1996, number 219 in Lecture Notes: 241 - 245.[5]Claypoole R L, Davis G, Sweldens W. Nonlinear wavelet transforms for image coding[R]. in Proc.31st Asilomar Conf,Pacific Grove, CA, 1997:301 - 308.[6]陈逢时.子波变换理论及其在信号处理中的应用[M].北京:国防工业出版社,1998:165-173.[7]Zou Hao-min. Wavelet transform and PDE techniques in image compression [D]. Doctor paper, University of California, 2000.[8]Daubechies I, Sweldens W. Factoring wavelet transforms into lifting steps [R]. Tech. Rep. Bell Laboratories, 1996.