Machine Learning in Medical Imaging

Miles N. Wernick

doi:10.1109/MSP.2010.936730

2010-07-01

250

DOI：10.1109/MSP.2010.936730

原文链接

查看原文

摘要

En 中文

Statistical methods of automated decision making and modeling have been invented ( and reinvented) in numerous fields for more than a century. Important problems in this arena include pattern classification, regression, control, system identification, and prediction. In recent years, these ideas have come to be recognized as examples of a unified concept known as machine learning, which is concerned with 1) the development of algorithms that quantify relationships within existing data and 2) the use of these identified patterns to make predictions based on new data. Optical character recognition, in which printed characters are identified automatically based on previous examples, is a classic engineering example of machine learning. But this article will discuss very different ways of using machine learning that may be less familiar, and we will demonstrate through examples the role of these concepts in medical imaging. Machine learning has seen an explosion of interest in modern computing settings such as business intelligence, detection of e-mail spam, and fraud and credit scoring. The medical imaging field has been slower to adopt modern machine-learning techniques to the degree seen in other fields. However, as computer power has grown, so has interest in employing advanced algorithms to facilitate our use of medical images and to enhance the information we can gain from them. Although the term machine learning is relatively recent, the ideas of machine learning have been applied to medical imaging for decades, perhaps most notably in the areas of computer-aided diagnosis (CAD) and functional brain mapping. We will not attempt in this brief article to survey the rich literature of this field. Instead our goals will be 1) to acquaint the reader with some modern techniques that are now staples of the machine-learning field and 2) to illustrate how these techniques can be employed in various ways in medical imaging using the following examples from our own research: CAD content-based image retrieval (CBIR) automated assessment of image quality brain mapping.

Keyword:

SUPPORT VECTOR MACHINE

PROCESSING PIPELINES

FMRI

CLASSIFICATION

STATE

CONNECTIVITY

SIMILARITY

PREDICTION

NPAIRS