Student Biomedical Computing Research Projects Daisy Y. Wong, (Ph.D. student in computer science) A Collaborative Interpreting System for Knowledge Discovery, University of Alabama at Birmingham Advisor: Dr. Warren Jones http://www.cis.uab.edu/info/grads/dw The last stage of a computerized knowledge discovery process, interpretation/evaluation, requires end users to make sense out of the patterns reported by data mining mechanisms from large databases. Currently, most knowledge discovery research concentrates on data mining and visualization techniques. Scant research has focussed on the interpretation of the mined output by domain experts. Organizations using current KDD tools have found a diverse team of data mining knowledgeable domain experts are needed to generate authoritative interpretations to support timely decision making. In most organizations, retaining such a team just for the purpose of data interpretation is neither feasible nor practical. In prior research at the University of Alabama at Birmingham, a Data Mining Surveillance System (DMSS) was developed for active infectious disease surveillance in a local hospital. That research also found that multiple domain experts were necessary to interpret the novel data mining output. Our current research specifically addresses the need for timely acquisition and compilation of interpretations by multiple experts who are distributed both geographically and temporally. We are designing a computer-based Collaborative Interpreting System (CIS) to support interpretation of output from data mining by a diverse group of domain experts collectively and on a continuous basis. Infectious disease surveillance is the problem domain used for testing the system. The aggregate interpretations of the experts will serve as an authoritative information source for infection control decision making. CIS is a web based client/server system to provide an interactive environment for collaborative interpreting. CIS uses DMSS as the datamining engine. The group decision support mechanism of CIS is adapted from a group communication process called the Delphi method. CIS will be used m examine the hypothesis that the results of collaboration are better than the result that could have been obtained by any single member of the group. This research i,s supported by a grant from the Alabama Academy of Science, 1998. Jeremy Aekerman, (M.D./Ph.D. student in biomedical engineering) AugmentedReality, Duke University Advisor: Dr. Henry Fuchs http://www.cs.unc.edu/-ackerman. Augmented reality combines computer generated graphics with the user's view of the real world. We are studying the application of this technique to medical problems. Modem medical techniques such as image guided and minimally invasive procedures increasingly rely on the clinician's ability to combine imagery from a variety of sources into a coherent view of an internal structure that is no longer directly visualized. Augmented reality potentially reduces the disadvantages of minimally invasive procedures by allowing the clinician to view the images in place inside the patient. This should allow more natural use of day-to-day hand-eye coordination than manipulating an instrument and observing the results from a spatial separate display. Successful application of augmented reality has a large number of technical challenges. A useful system must operate in real-time with a high frame rate leading large graphics generation and processing requirements. Displays for augmented reality are not widely available and frequently do not meet the demands of medical applications. Tracking devices used for augmented reality must be extremely fast and high precision so that synthetic imagery is properly registered with the real world. Finally, imaging devices must be calibrated or developed that supply the necessary information for creating useful graphics. Our research has focused on two specific applications of augmented reality: ultrasound guided needle breast biopsy and laparoscopic surgery.

/lp/association-for-computing-machinery/student-biomedical-computing-research-projects-K0MBeZpNxP