知识与财富的链接
有效的医学图像增强技术应将感兴趣目标或区域增强、背景抑制和噪声削减综合考虑,能改善图像的质量,在减少噪声的同时保持着原有的纹理特征,有助于后续得到正确的临床诊断结果,这对某些疾病的早期确诊有极大帮助。归纳了基于非局部信息的医学图像增强技术常见方法,包括非局部均值滤波算法、三维块匹配算法、形态成分分析算法等,通过介绍这几种方法原理,指出这些方法的使用范围及现状,并进行了性能对比分析,最后探索性地给出了现阶段医学图像增强技术的可能发展方向之一,即基于上下文量化的图像增强技术。
Taking into account the improvement of the contrast between regions of interest and the background and denoising, the effective techniques of medical image enhancement can improve image quality and maintain the original characteristics of the texture while denoising, which is helpful to get the correct results in the early clinical diagnosis of disease. This paper surveyed several common methods based on non-local patch medical image enhancement, including non-local mean filtering algorithm, three-dimensional block-matching (BM3D) algorithm, and morphological component analysis (MCA) algorithm. And it described and compared the status and principle of these methods. Finally, it introduced one of the currenty trends for medical image enhancement, which was based on the context quantization.
