Benefits of Three Dimensional Data in Characterizing Images

点击翻译按钮获取中文摘要

This research project is concerned with ultrasonic tissue characterization (UTC) which is an automated means of detecting and classifying diseased tissue using an ultrasound image. Current research in this area is based upon computer manipulation of two- dimensional (2-D) ultrasound images. This project will extend current 2-D to 3-D images. Use of 3-D images permits the characterization of tissue texture along any plane or using any volume. Plane will be defined by vectors or by some boundary in the 3-D image, i.e., the boundary between the kidney wall and the liver wall. Existing 2-D algorithms will be transformed to the 3-D format. Devising a general mapping of any 2-D algorithm is complex, but fortunately the particular parameters involved lend themselves to this transformation. The ultimate goal is to present a 3-D color-coded image to a human operator for interpretation, manipulation and diagnosis. This combination of machine computation and human intelligence leads to increased accuracy and has the potential of discovering new techniques.