A curated list of resources for 3D segmentation of neurites (connectomics) in EM images

Papers

• 2017-BioInf - DeepEM3D: approaching human-level performance on 3D anisotropic EM image segmentation. [Paper][Code]

• 2017-Arxiv - Superhuman Accuracy on the SNEMI3D Connectomics Challenge. [Paper][Code]

• 2017-NIPS - An Error Detection and Correction Framework for Connectomics. [Paper]

• 2017-ICCVW - Solving large Multicut problems for connectomics via domain decomposition. [Paper]

• 2017-NatureMethods - Multicut brings automated neurite segmentation closer to human performance. [Paper]

• 2018-BMVC - Efficient Correction for EM Connectomics with Skeletal Representation. [Paper]

• 2018-NatureMethods - High-Precision Automated Reconstruction of Neurons with Flood-filling Networks. [Paper][Arxiv][Code]

• 2018-FNC - Analyzing Image Segmentation for Connectomics. [Paper]

• 2019-CVPR - Cross-Classification Clustering: An Efficient Multi-Object Tracking Technique for 3-D Instance Segmentation in Connectomics. [Paper]

• 2019-CVPR - Biologically-Constrained Graphs for Global Connectomics Reconstruction. [Paper]

• 2019-CVPR - End-to-End Learned Random Walker for Seeded Image Segmentation. [Paper]

• 2019-ISBI - Learning Metric Graphs for Neuron Segmentation In Electron Microscopy Images. [Paper]

• 2019-ISBI - Reconstructing neuronal anatomy from whole-brain images. [Paper]

• 2019-CON - Convolutional nets for reconstructing neural circuits from brain images acquired by serial section electron microscopy. [Paper]

• 2019-CON - Big data in nanoscale connectomics, and the greed for training labels. [Paper]

• 2019-TPAMI - Large Scale Image Segmentation with Structured Loss Based Deep Learning for Connectome Reconstruction. [Paper][Arxiv][Code]

• 2019-MM - Automated reconstruction of a serial-section EM Drosophila brain with flood-filling networks and local realignment. [Paper]

• 2019-SR - UNI-EM: An Environment for Deep Neural Network-Based Automated Segmentation of Neuronal Electron Microscopic Images. [Paper]

• 2019 - Entropy policy for supervoxel agglomeration of neurite segmentation. [Paper]

• 2019 - A Generalized Framework for Agglomerative Clustering of Signed Graphs applied to Instance Segmentation. [Paper]

• 2019 - Reconstructing neurons from serial section electron microscopy images. [Thesis]

• 2020 - Robust neural circuit reconstruction from serial electron microscopy with convolutional recurrent networks. [Paper]

• 2020 - Optimizing the Computational Efficiency of 3D Segmentation Models for Connectomics. [Paper]

• 2020 - Machine Learning for Instance Segmentation. [Thesis]

• 2020 - Generative and discriminative model-based approaches to microscopic image restoration and segmentation. [Paper]

• 2020 - Accelerated EM Connectome Reconstruction using 3D Visualization and Segmentation Graphs. [Paper]

• 2020 - Machine Learning for Connectomics. [Thesis]

• 2020-TPAMI - The Mutex Watershed and its Objective: Efficient, Parameter-Free Graph Partitioning. [Paper][Code]

• 2021-TMI - Learning Dense Voxel Embeddings for 3D Neuron Reconstruction. [Paper][Code]

• 2021 - Scalable Instance Segmentation for Microscopy. [Thesis]

• 2022-TMI - Semi-Supervised Neuron Segmentation via Reinforced Consistency Learning. [Paper][Code]

• 2023-NatureMethods - Local Shape Descriptors for Neuron Segmentation. [Paper][Code]

Benchmark Datasets

• 2013-ISBI - SNEMI3D: [Dataset & Challenge details]

• 2015-PNAS - FIB-25: Synaptic circuits and their variations within different columns in the visual system of Drosophila. [Paper] [Dataset & details]

• 2015-Neuron - SegEM: SegEM: Efficient Image Analysis for High-Resolution Connectomics. [Paper][Dataset & Challenge details]

• 2016-MICCAI - CREMI: [Dataset & Challenge details]

• 2018-Arxiv - star-challenge: Robust neural circuit reconstruction from serial electron microscopy with convolutional recurrent networks. [Paper][Dataset & Challenge details]

Metrics

• 2013-ISBI - Adapted Rand error: SNEMI3D [Definition][Matlab-Code][Python-code]

• 2015-PNAS - Inter-error distance (IED): SegEM: Efficient Image Analysis for High-Resolution Connectomics. [Definition][Matlab-Code]

• 2015-Methods - Tolerant Edit Distance (TED): TED: A Tolerant Edit Distance for Segmentation Evaluation. [Paper][Code]

• 2016-MICCAI - Variation of Information (VOI): CREMI: [Links] [Python-code]

• 2018-NatureMethods - Expected Run Length (ERL): High-Precision Automated Reconstruction of Neurons with Flood-filling Networks. [Definition]

• 2018-FN - Neural Reconstruction Integrity (NRI): A Metric for Assessing the Connectivity Accuracy of Reconstructed Neural Networks. [Paper]

• 2019-NeurIPS - Betti number error: Topology-Preserving Deep Image Segmentation. [Paper]

• 2023-NatureMethods - Min-Cut Metric (MCM): Local Shape Descriptors for Neuron Segmentation. [Paper][Code]

Other Resources

• PyTorch Connectomics - [Code]

• pytorch-3dunet - [Code]

• Awesome-vEM-Datasets - [Dataset details]