OBJECT DEFINITION IN TOMOGRAPHIC RADIOLOGY

University Of Pennsylvania, Philadelphia PA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (Provided by Applicant): The broad long-term goal of this project is to develop theories, algorithms and their practical computer implementations to identify and quantify object information captured in multidimensional medical images. In many radiological applications, the lack of cost-effective methods for this purpose with proven, acceptable precision and accuracy remains one of the major impediments to further advances. With this in mind, this proposal addresses three goals: (i) to standardize the MR image intensity scale so that object definition and tissue characterization is facilitated; (ii) to segment tissue regions in the brain into anatomic regions and into diseased tissue regions; (iii) to devise new MR image-based quantitative measures that would give more disease-specific information in Multiple Sclerosis (MS). The standardization method is based entirely on image processing (histogram deformation), rather than on phantoms, and hence it will be applicable to already acquired images also. The segmentation of tissue regions is done via an extension of the fuzzy connectedness method to multiple objects wherein it considers the relative connectedness among objects to delineate their fuzzy interfaces. The new measures to quantify the MS disease severity are based on the standardized image intensity distributions within tissue regions. MS is an acquired disease of the central nervous system whose cost to the US is estimated at $2.5 billion annually. A precise, accurate, cost-effective image-based method of determining the disease severity is currently lacking. The proposed methods may not only help in understanding MS and its progression but also in determining the effectiveness of various therapies.