Faculty of Engineering and Natural Sciences
Image Processing Challenges in Neural Circuit Reconstruction
Tolga Taşdizen, University of Utah
In this talk, we will focus on the problem of automated neural circuit reconstruction as the primary application driving our image processing and pattern recognition research. Neural circuit reconstruction, from simple organisms to the human brain, is an area of research that offers great promise in understanding and modeling animal behavior. Serial-section electron microscopy images can provide the data necessary for reconstruction of such large-scale neural circuits. However, the complexity and vast size of these images make human interpretation an extremely labor intensive task for all but the smallest circuits. The pipeline for reconstructing neural circuits from serial-section electron microscopy includes preprocessing the images, assembling 3D volumes, segmenting individual neurons and identifying synapses between the neurons. We will discuss our efforts in the automation of this pipeline. In the second part of the talk, we will focus on two technical research problems: an optimal path finding approach for neuron segmentation from highly anisotropic data and a nonlinear dimensionality reduction algorithm for working with high-dimensional image features.
Biography: Dr. Taşdizen received his B.S. degree from Boğaziçi University in 1995, M.Sc. and Ph.D. degrees from Brown University in 1997 and 2001 respectively. He has been with University of Utah since 2001. His research interests include image processing, pattern recognition, applications in biological image analysis and classification problems in medical image analysis.
May 26, 2008, 10:00, FENS L056