Superhuman Accuracy on the SNEMI3D Connectomics Challenge

Abstract

For the past decade, convolutional networks have been used for 3D reconstruction of neurons from electron microscopic (EM) brain images. Recent years have seen great improvements in accuracy, as evidenced by submissions to the SNEMI3D benchmark challenge. Here we report the first submission to surpass the estimate of human accuracy provided by the SNEMI3D leaderboard. A variant of 3D U-Net is trained on a primary task of predicting affinities between nearest neighbor voxels, and an auxiliary task of predicting long-range affinities. The training data is augmented by simulated image defects. The nearest neighbor affinities are used to create an oversegmentation, and then supervoxels are greedily agglomerated based on mean affinity. The resulting SNEMI3D score exceeds the estimate of human accuracy by a large margin. While one should be cautious about extrapolating from the SNEMI3D benchmark to real-world accuracy of large-scale neural circuit reconstruction, our result inspires optimism that the goal of full automation may be realizable in the future.

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Natural Scales in Geographical Patterns
• Deep Residual Learning for Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• FusionNet: A deep fully residual convolutional neural network for image segmentation in connectomics
• Image Segmentation by Size-Dependent Single Linkage Clustering of a Watershed Basin Graph
• Learned versus Hand-Designed Feature Representations for 3d Agglomeration

Cited by in corpus (36)

• Medical Image Segmentation Using Deep Learning: A Survey
• MIScnn: A Framework for Medical Image Segmentation with Convolutional Neural Networks and Deep Learning
• Review: Deep Learning in Electron Microscopy
• Segmentation in large-scale cellular electron microscopy with deep learning: A literature survey
• The Mutex Watershed and its Objective: Efficient, Parameter-Free Graph Partitioning
• Improving Electron Micrograph Signal-to-Noise with an Atrous Convolutional Encoder-Decoder
• Self-Supervised Neuron Segmentation with Multi-Agent Reinforcement Learning
• A Hierarchical Probabilistic U-Net for Modeling Multi-Scale Ambiguities
• Automatic Volumetric Segmentation of Additive Manufacturing Defects with 3D U-Net
• Deep Multimodal Fusion of Data with Heterogeneous Dimensionality via Projective Networks
• Finding Nano-Ötzi: Semi-Supervised Volume Visualization for Cryo-Electron Tomography
• Chunkflow: Distributed Hybrid Cloud Processing of Large 3D Images by Convolutional Nets
• An Error Detection and Correction Framework for Connectomics
• Large-scale image segmentation based on distributed clustering algorithms
• Deep Residual 3D U-Net for Joint Segmentation and Texture Classification of Nodules in Lung
• HadaNets: Flexible Quantization Strategies for Neural Networks
• Non-local U-Net for Biomedical Image Segmentation
• Robust neural circuit reconstruction from serial electron microscopy with convolutional recurrent networks
• Learning and Segmenting Dense Voxel Embeddings for 3D Neuron Reconstruction
• Self-supervised learning via inter-modal reconstruction and feature projection networks for label-efficient 3D-to-2D segmentation
• Scan2Part: Fine-grained and Hierarchical Part-level Understanding of Real-World 3D Scans
• Morphomics via Next-generation Electron Microscopy
• Seeing eye-to-eye? A comparison of object recognition performance in humans and deep convolutional neural networks under image manipulation
• Detecting Synapse Location and Connectivity by Signed Proximity Estimation and Pruning with Deep Nets
• Contrastive Learning for Mitochondria Segmentation
• NucMM Dataset: 3D Neuronal Nuclei Instance Segmentation at Sub-Cubic Millimeter Scale
• PZnet: Efficient 3D ConvNet Inference on Manycore CPUs
• Scaling Distributed Training of Flood-Filling Networks on HPC Infrastructure for Brain Mapping
• End-to-End Learned Random Walker for Seeded Image Segmentation
• Stateless actor-critic for instance segmentation with high-level priors
• Voxel-level Siamese Representation Learning for Abdominal Multi-Organ Segmentation
• Convolutional nets for reconstructing neural circuits from brain images acquired by serial section electron microscopy
• Sub-cortical structure segmentation database for young population
• Assisting human experts in the interpretation of their visual process: A case study on assessing copper surface adhesive potency
• AxonEM Dataset: 3D Axon Instance Segmentation of Brain Cortical Regions
• Consistent Recurrent Neural Networks for 3D Neuron Segmentation