小波域基于层内和层间系数约束的边缘软判决与图像去噪

在提出依概率对边缘系数软判决的基础上，运用图像小波分解后在尺度和空间方向的相关性，引入调谐函数，对边缘发生概率增加邻域约束，以进一步提高边缘发生概率的估计精度。应用于图像去噪的仿真结果表明，引入层问和层内约束后，在较好滤除噪声的同时，图像边缘也得到了明显加强。该方法也可用于边缘检测和图像分割。

Wavelet Based Edge Soft-decision Using Intrascale and Interscale Coefficients Constraints for Image Denoising

An approach of soft-decision to edge is firstly proposed, based on estimation of edge probability. Moreover, the harmonic functions are respectively applied to scale and spatial neighbor coefficients as constraint in order to improve the estimation precision of probability, by using the correlations of wavelets coefficients in scale and space directions. The application results of image denoising show the image edge is implicitly preserved, whilst image noise is filtered out, after using intrascale and interscale constraints. The proposed approach also can be used in image edge detection, and image segmentation.