Computerized decision support in medical imaging

Abstract

Describes challenges in using image processing and automated feature extraction for improving diagnostic accuracy. The generalization of digital technology in medical imaging makes it possible to design computer-aided decision-support tools to help resolve the difficulties encountered by radiologists. It has already been demonstrated that image processing and automated feature extraction can help improve diagnostic accuracy in some applications. Future challenges are for the development of vision techniques tailored for a wider range of imaging modalities and pathologies the design of methods for the combination of several classifiers, and the implementation of expert systems incorporating both low-level feature extractors and high-level clinical knowledge. After evaluation, these tools will be integrated in PACS systems to highlight regions of interest and to provide clinicians with a documented second opinion. Then, such integrated systems can be used routinely in the clinical practice and teaching of radiology.

Ähnliche Arbeiten