Computer Science Research

Fetal Alcohol Syndrome (FAS) Diagnosis

Fetal Alcohol Syndrome (FAS) Disorder is a common alcohol related neurological disorder caused by alcohol exposure during early pregnancy. Children with FAS disorder exhibit an array of cognitive, behavioral, and emotional deficits that impair daily functioning in many domains. FAS also exhibit characteristic facial features that are often used in FAS diagnosis. Recently, NIH issued a solicitation for a Consortium for the "Collaborative Initiative on Fetal Alcohol Spectrum Disorders" to develop effective early diagnosis and treatment solutions. Using a 3D laser scanner, we are able to gather detailed 3D facial data (geometry and color) for children of different ages, races and origins with and without FAS. The ability to automatically collect, measure and analyze 3D facial features is very novel in medical diagnosis, and offers unlimited potential for a wide range of future medical solutions. The data analysis process consists of two stages:

1. using currently known FAS related facial features for automatic early diagnosis. This requires 3D interactive visualization, automatic measurements, and 3D surface level computation.
2. 3D feature analysis at both geometry and texture levels, and machine learning based on case studies. Current diagnosis criteria are very primitive and unreliable. Quantitative and qualitative analysis provides not only accurate feature measures, but also a way of discovering new and more complex diagnosis features that are impossible to determine manually. As part of an NIH proposal, VISC is working closely with a group from IU Medical Center as well as several other groups nationawide on a collaborative initiative for FAS diagnosis.

Project Presented By

Fang, Shiaofen

Position:	Chair, Professor
Mail:
Phone:	317-274-9731
Url:	http://www.cs.iupui.edu/~sfang

Education Details

Research Interests

My primary research interests are in visualization, biomedical imaging, computer graphics, and geometric modeling. My early research focus had been on Volume Visualization and Volume Graphics. Major projects include deformable volume modeling and rendering, hardware assisted voxelization, volume fusion, 3D microscopy visualization, and immersive volume visualzation. A common theme of these research projects is the efficient and effective visaulization of scientific data. Although this line of research is still a major part of my work, My recent research interests have been shifted more towards the direction of medical image analysis and visual data mining. These include 3D image analysis for medical diagnosis, surface analysis of medical scans, 3D surface reconstruction, and knowledge discovery through information visualization.