Consistent Recurrent Neural Networks for 3D Neuron Segmentation

TL;DR

This paper introduces a recurrent neural network that ensures spatio-temporal consistency in 3D neuron segmentation, producing accurate object masks without post-processing, and achieves state-of-the-art results on benchmark datasets.

Contribution

The paper proposes a novel end-to-end recurrent network that models local and non-local object relationships for 3D neuron segmentation, improving over boundary-based methods.

Findings

Achieved state-of-the-art performance on SNEMI3D benchmark.

Produces consistent 3D neuron masks without post-processing.

Effectively models local and long-range object relationships.

Abstract

We present a recurrent network for the 3D reconstruction of neurons that sequentially generates binary masks for every object in an image with spatio-temporal consistency. Our network models consistency in two parts: (i) local, which allows exploring non-occluding and temporally-adjacent object relationships with bi-directional recurrence. (ii) non-local, which allows exploring long-range object relationships in the temporal domain with skip connections. Our proposed network is end-to-end trainable from an input image to a sequence of object masks, and, compared to methods relying on object boundaries, its output does not require post-processing. We evaluate our method on three benchmarks for neuron segmentation and achieved state-of-the-art performance on the SNEMI3D challenge.