The Relevance of Computer-Aided-Diagnosis Systems in Microscopy Applications to Medicine and Biology

The Relevance of Computer-Aided-Diagnosis Systems in Microscopy Applications to Medicine and Biology

Paolo Soda (Università Campus Bio-Medico di Roma, Italy) and Giulio Iannello (Università Campus Bio-Medico di Roma, Italy)
Copyright: © 2008 |Pages: 8
DOI: 10.4018/978-1-59904-889-5.ch147
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In this article, we discuss the use of computerbased systems in microscopy, focusing on cytological images. We initially present recent results on image segmentation, and then we argue that it makes sense moving from a structural approach to a semantic interpretation of micrographs. In this respect, we focus on the relevance of using CAD tools to overcome the current limitations of microscopy, investigating several peculiar objectives of such systems. A short review of the literature demonstrates that the development of a flexible CAD applicable to various working scenarios is a future trend in microscopy healthcare systems. To support our position, we briefly describe a tool that analyzes and classifies fluorescence images.

Key Terms in this Chapter

Micrographs: Microscope images.

Features: A set of attributes that represents a pattern.

Image Segmentation: Partition of an image into not overlapping, constituent regions that are homogeneous with respect to some characteristic. Ideally, a segmentation method finds sets of points that correspond to distinct structures or regions of interest in the image.

Multi-Expert System (MES): Aggregation of classification systems according to a combination approach, in order to improve the recognition performance.

Computer-Aided Diagnosis (CAD) System: Computer-based system that analyzes medical data in order to support the human specialist in the diagnosis process.

Reject Option: Classification option that aims at rejecting the highest possible number of samples that would otherwise be misclassified (i.e., misclassified without a reject option). It is usually based on the estimation of the classification reliability.

Image Classification: Recognition the patterns or the objects of an image.

Complete Chapter List

Search this Book:
Reset