A Multi-Pass Approach to Large-Scale Connectomics

TL;DR

This paper introduces a multi-pass, real-time pipeline for large-scale connectomics data processing, significantly speeding up neuronal connectivity reconstruction from electron microscopy images using novel neural network methods.

Contribution

It presents a multi-pass approach optimized for multicore systems, enabling near real-time processing of petabyte-scale neural imaging data, with novel neural nets and error detection methods.

Findings

Reconstructed a large neural dataset in hours instead of weeks.

Developed a fast-pass neural network for initial segmentation.

Demonstrated error correction with a slow-pass refinement.

Abstract

The field of connectomics faces unprecedented "big data" challenges. To reconstruct neuronal connectivity, automated pixel-level segmentation is required for petabytes of streaming electron microscopy data. Existing algorithms provide relatively good accuracy but are unacceptably slow, and would require years to extract connectivity graphs from even a single cubic millimeter of neural tissue. Here we present a viable real-time solution, a multi-pass pipeline optimized for shared-memory multicore systems, capable of processing data at near the terabyte-per-hour pace of multi-beam electron microscopes. The pipeline makes an initial fast-pass over the data, and then makes a second slow-pass to iteratively correct errors in the output of the fast-pass. We demonstrate the accuracy of a sparse slow-pass reconstruction algorithm and suggest new methods for detecting morphological errors. Our fast-pass approach provided many algorithmic challenges, including the design and implementation of novel shallow convolutional neural nets and the parallelization of watershed and object-merging techniques. We use it to reconstruct, from image stack to skeletons, the full dataset of Kasthuri et al. (463 GB capturing 120,000 cubic microns) in a matter of hours on a single multicore machine rather than the weeks it has taken in the past on much larger distributed systems.