Carnegie Mellon University
Advanced Search   
  Look in
       Title     Description
       Inactive Projects
Biomedical Image Analysis
This lab is no longer active.
Head: Yanxi Liu
Contact: Yanxi Liu
Associated center(s) / consortia:
 Medical Robotics Technology Center (MRTC)
 Vision and Autonomous Systems Center (VASC)
The long-term objective of our research is to build a computational framework for automatic disease classification, discrimination and prediction. We take an image feature-based statistical multivariate machine learning approach on multimodal biomedical images including, but not limited to, high resolution Magnetic Resonance Images (MRI), CT images, multispectral microscopic images and optical photos and videos. Working closely in team composed of computer scientists, neuropsychologist, geriatric psychiatrist, neurologist, neuroradiologist and psychologist, we have a wide range of applications with one focused goal: discovering the discriminative feature subspaces for automatic object semantic class prediction. To reach this goal, our work covers the development of computer algorithms for learning-based deformable registration, atlas-based segmentation, 3D shape representation and analysis, innovative image feature extraction and discriminative feature subspace induction and selection. We have applied our method successfully in CT neural images for discriminating among normal, infarct and blood cases for image content-based retrieval from large, multimedia stroke patient databases ; hyper-spectral Pap smear microscopic images for screening cancer cells from normal cells; facial expression videos for human identification and even digital videos for real-time moving target tracking .Our current challenging and exciting projects include automatic classifcation of neuropsychiatric patients with central nervous system (CNS) diseases, currently focusing on schizophrenia or Alzheimer's disease, from healthy subjects using high resolution Magnetic Resonance Images (MRI).