多尺度生物医学数据可视化

项目名称： 多尺度生物医学数据可视化

项目编号： No.61472354

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 计算机科学学科

项目作者： 陶煜波

作者单位： 浙江大学

项目金额： 80万元

中文摘要： 生命系统不同尺度的器官、组织和细胞等，存在一定程度的功能关联性，综合不同尺度数据的可视化能够全面理解相关组织的功能及其相互协作关系。多尺度生物医学数据可视化是多模态数据可视化从相似尺度到不同尺度的扩展，尚存在以下问题：尺度的不连续性，多尺度特征分析能力不强，多尺度数据融合机制不完善。针对这些问题，本项目将深入研究多尺度数据的表示、分析和融合方法，揭示从多尺度数据中快速高效获取隐含信息的可视化机制。具体包括：(1) 构建连续尺度数据表示，实现多尺度数据无缝缩放；(2) 综合属性、图像、数据和特征空间的相似性和差异性交互式探索分析；(3) 基于特征分析的数据滤波、可视化映射和体绘制积分累加阶段的数据融合；(4) 基于形状连续性的时变特征提取，构建特征状态转换图，分析多尺度结构协作关系。本项目还将提出便于分析多尺度数据的用户交互模式，实现集成上述技术的多尺度数据可视化原型系统，验证所提出的方法。

中文关键词： 体数据；多尺度；连续表示；特征分析；数据融合

英文摘要： Multiscale data, such as organ, tissue, cell, et al in the life system, have functional correlations among them, and visualizing these data integratively can assist to fully understand the functionalities of different parts and their cooperation relationships. Multiscale biomedical visualization is extended from similar scales in multi-modal visualization to different scales, and it still has several unsolved problems: discontinuity in the scale dimension, inaccurate enough of multiscale feature analysis, and incomplete multiscale data fusion schemes. In order to address these problems, this project proposes to conduct an in-depth study on the representation, analysis and fusion techniques of multiscale data, and to develop novel algorithms for exploiting information inside multiscale data effectively. The specific contents include: (1) construct a data representation in the continuous scale, and implement seamless scaling of multiscale data; (2) integrate the attribute, image, data and feature exploratory spaces, and interactively analyze similarities and differences in multiscale data; (3) fuse multiscale data in the data filtering, visualization mapping and accumulation in the rendering stage based on feature analysis; (4) extract time-varying features based on shape coherence in time, construct state-transition graphs of time-varying features, and analyze the interplay of multiscale spatial patterns in time. This project also plans to design a new user interface for the effective analysis of multiscale data, and to implement the visualization system, which integrates the above techniques for multiscale data, to demonstrate the effectiveness of the proposed methods.

英文关键词： volume data;multiscale;continuous representation;feature analysis;data fusion