Clinical Informatics
A Scalable Platform for Exploring and Analyzing Whole Brain Tissue Cleared Images
Description
The ability to accurately localize and characterize cells in light sheet fluorescence microscopy (LSFM) images is indispensable for shedding new light on the understanding of three dimensional structures of the whole brain. We are designing a semi-automatic annotation workflow to largely reduce human intervention, and thus improve both the accuracy and the replicability of annotation across different users. The annotation software will be expanded into a crowd-sourcing platform which allows us to obtain a massive number of manual annotations in a short time. We will also develop a fully 3D cell segmentation engine using 3D convolutional neural networks trained with the 3D annotated samples. We will then develop a transfer learning framework to make our 3D cell segmentation engine general enough for the application of novel LSFM data which might have a significant gap of image appearance due to different imaging setup or clearing/staining protocol. This general framework will allow us to rapidly develop a specific cell segmentation solution for the new LSFM data with very few or even no manual annotations, by transferring the well-trained segmentation engine that has been trained with a sufficient number of labeled samples. Finally, we will apply our computational tool to several pilot studies including Autism mouse brain and human fetal tissue data.
RENCI's Role
RENCI has developed a standalone application (Segmentor) and a crowd-sourcing platform for annotation (ninjato) that will generate high-quality training data for a 3D cell segmentation engine capable of annotating new microscopy images automatically. Collaborators from the Stein lab in the UNC Department of Genetics will apply this computational tool to several pilot studies including Autism mouse brain and human fetal tissue data.