||To make Quantitative Radiology a reality in clinical practice and to exploit the benefits of BigData for improving clinical care, computerized body-wide automatic anatomy recognition (AAR) in medical images becomes essential. With the goal of building a general AAR system that is not tied to any specific organ system, body region, or image modality, this presentation will describe an AAR methodology developed over the past 8 years for localizing and delineating all major objects in different body regions. The methodology embodies the following key ideas: (a) Exploiting the large collection of existing patient images. (b) Formulating precise anatomic definition of each body region and all its major objects and delineating them following these definitions. (c) Building hierarchical fuzzy anatomy models of object assemblies body-wide by fully exploiting detailed knowledge of the form, size, and positional relationships of objects, by separating the modality-specific and independent parts, and by modeling sparse and non-sparse objects differently. (d) Recognizing (locating) objects in given images by employing the hierarchical anatomy models and using an optimal threshold-based search. (e) Delineating objects following the hierarchical order via fuzzy model-based fuzzy connectedness principles.
The AAR system has been tested on four body regions – neck (on MRI and CT), thorax (on CT and PET), abdomen (on CT, MRI, and PET), and pelvis (on CT and PET) – involving a total of about 50 objects. Overall the AAR method achieves a mean accuracy of 1-3 voxels in localizing (recognizing) non-sparse and sparse objects and delineation accuracy expressed as mean false positive and negative volume fractions of 2%-8%, with mean boundary distance relative to ground truth of 1-2 voxels. Separation of modality-independent from dependent aspects, organization of objects in a hierarchy, encoding of object relationship information explicitly into the hierarchy, optimal threshold-based recognition learning, differently modeling sparse and non-sparse objects, and fuzzy model-based fuzzy connectedness are effective concepts which resulted in a general AAR system that works in different body regions, on a variety of organs, on different modalities, and for a variety of applications.
Biography: Jayaram Udupa received a bachelor’s degree in Electronics and Communication Engineering from Mysore University, India, in 1972, and a PhD in Computer Science from the Indian Institute of Science, Bangalore, in 1976. From the beginning of his career, his research focus has been developing theory, algorithms, and large software systems for image processing, 3D visualization, image analysis, and in utilizing these in numerous medical application areas toward quantitative radiology. He has made numerous seminal contributions to these areas continuously for nearly 40 years and contributed significantly to the initiation and success of several conferences such as the SPIE Medical Imaging Symposium and MICCAI since the early days of medical image processing. He has published 190 journal papers, 217 full conference papers, 2 books, 25 book chapters, given 275 invited lectures worldwide, and trained about 70 PhD students and post-doctoral fellows. He is a Fellow of the IEEE and American Institute of Medical and Biological Engineering (AIMBE), a Professor of Radiological Sciences and Chief of the Medical Image Processing Section, Dept. of Radiology, University of Pennsylvania, Philadelphia. In addition, he has visited Uppsala University several times before over the years.