Speaker: Jingpeng Wu, PhD
Title: Towards Petabyte Scale Neuron Reconstruction using Deep Learning and Cloud Computing
The computation of neural networks is closely related to its connectivity and morphology. Recently, advanced imaging methods with both large field of view and high resolution have acquired this information and produced large scale image stacks. However, it is still challenging to reconstruct neurons from these images. In order to reconstruct neurons from Serial Section Electron Microscopy images, we have developed a pipeline based on Deep Learning and Cloud Computing. I’ll use a zebrafish dataset to showcase our pipeline focusing on my contributed parts. Then, I’ll also show some reconstruction results from other datasets. At last, I’ll talk about the potential application to Light Microscopy images.
Two datasets are publicly available. Jingpeng Wu contributed to the neuron reconstruction of them and will be mention them in the talk.
Jingpeng Wu is a Postdoctoral Research Associate at the Seung Lab in Princeton Neuroscience Institute. His research interests include Connectome, 3D image processing, neuron reconstruction.