Hybrid Pipelining Approach to Image Alignment for Large-Scale Brain Image Data

R. Sengupta, J. Meyer, and Z. Zhang (USA)


Image Registration, Large-Scale Data Visualization, Histograms, Thresholding, Feature Extraction


Creating thin sections of frozen tissue in the order of a few microns and then manually mounting the resulting slices on glass plates is a common technique in brain imaging. Large collections of such manually mounted cryosections are available for digital archiving. Scanning these slices at high resolution helps preserving them for future generations. The obtained data can also be used to restore the original shape of the specimen, i.e., to create a three-dimensional model. This task usually requires time-consuming individual alignment of the slices. We present a framework that uses a pipelining approach to aid in the alignment process of a large data set, and to automate most of the steps.

Important Links:

Go Back