Statistical Data Mining for Medical Image Data Analysis

本项目的目的是开发新的统计数据挖掘模型和方法，用于医学图像数据分析。例如，扩散张量成像（DTI）是一种先进的磁共振成像（MRI）方法，使临床医生能够监测人类大脑疾病的进展，规划神经外科手术和研究大脑发育。DTI数据分析必须考虑具有复杂数学结构的扩散张量、图像的三维空间特征以及大脑中复杂纤维结构的存在。这些方面的结合使得扩散张量图像分析成为医学图像分析领域中最具挑战性的课题之一。另一个例子是乳房X线摄影数据分析。识别乳腺X线图像中的肿块、钙化、结构畸变等特征并分析其几何特征是乳腺癌筛查的关键。为了实现本项目的目标，将在快速发展的统计数据挖掘理论以及统计形状分析和计算机视觉的基础上开发新技术。

The aim of this project is to develop new statistical data mining models and methods for medical image data analysis. For example, Diffusion Tensor Imaging (DTI) is an advanced magnetic resonance imaging (MRI) method that benefits clinicians in monitoring human brain disease progression, planning neurosurgery and studying brain development. DTI data analysis must account for the diffusion tensor, which has a complicated mathematical structure, the three-dimensional spatial characteristics of the image, and the existence of complicated fibre structures in the brain. The combination of these aspects makes diffusion tensor image analysis one of the most challenging topics in the area of medical image analysis. Another example is mammography data analysis. To identify mammographic feature (e.g. mass, calcifications and architecture distortion) and to analyse the geometry properties of these features among mammography images are essential for breast cancer screening. To achieve the aim of this project, new techniques will be developed building on the rapidly developing theory of Statistical Date Mining, as well as on Statistical Shape Analysis and Computer Vision.