SET OPERATOR-BASED METHOD OF DENOISING MEDICAL VOLUME DATA

• 作者: 程兵 袁泽剑 郑南宁
• 来源期刊: 西安交通大学学报：英文版
• 学科: 工学
• 关键词: volume RENDERING SET OPERATOR VISUALIZATION noise SUPPRESSION level SET
• 年，卷（期）: 2002,（2）
• 页码范围: 109-112
• 页数: 4页
• 分类号: TP391.41

摘要

Objective To investigate impulsive noise suppression of medical volume data. Methods The volume data is represented as level sets and a special set operator is defined and applied to filtering it. The small connected components, which are likely to be produced by impulsive noise, are eliminated after the filtering process. A fast algorithm that uses a heap data structure is also designed. Results Compared with traditional linear filters such as a Gaussian filter, this method preserves the fine structure features of the medical volume data while removing noise, and the fast algorithm developed by us reduces memory consumption and improves computing efficiency. The experimental results given illustrate the efficiency of the method and the fast algorithm. Conclusion The set operator-based method shows outstanding denoising properties in our experiment, especially for impulsive noise. The method has a wide variety of applications in the areas of volume visualization and high dimensional data processing.