On the Compactness, Efficiency, and Representation of 3D Convolutional Networks: Brain Parcellation as a Pretext Task

Abstract

Deep convolutional neural networks are powerful tools for learning visual representations from images. However, designing efficient deep architectures to analyse volumetric medical images remains challenging. This work investigates efficient and flexible elements of modern convolutional networks such as dilated convolution and residual connection. With these essential building blocks, we propose a high-resolution, compact convolutional network for volumetric image segmentation. To illustrate its efficiency of learning 3D representation from large-scale image data, the proposed network is validated with the challenging task of parcellating 155 neuroanatomical structures from brain MR images. Our experiments show that the proposed network architecture compares favourably with state-of-the-art volumetric segmentation networks while being an order of magnitude more compact. We consider the brain parcellation task as a pretext task for volumetric image segmentation; our trained network potentially provides a good starting point for transfer learning. Additionally, we show the feasibility of voxel-level uncertainty estimation using a sampling approximation through dropout.

Paper accepted at IPMI 2017

References in corpus (5)

• Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
• What makes ImageNet good for transfer learning?

Cited by in corpus (61)

• Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations
• Interactive Medical Image Segmentation using Deep Learning with Image-specific Fine-tuning
• CA-Net: Comprehensive Attention Convolutional Neural Networks for Explainable Medical Image Segmentation
• Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks
• TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning
• NiftyNet: a deep-learning platform for medical imaging
• Automatic Brain Tumor Segmentation using Cascaded Anisotropic Convolutional Neural Networks
• DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation
• Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation
• An application of cascaded 3D fully convolutional networks for medical image segmentation
• Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities and Results of the WMH Segmentation Challenge
• Cross-Modality Deep Feature Learning for Brain Tumor Segmentation
• UNETR: Transformers for 3D Medical Image Segmentation
• CAI4CAI: The Rise of Contextual Artificial Intelligence in Computer Assisted Interventions
• Generalised Wasserstein Dice Score for Imbalanced Multi-class Segmentation using Holistic Convolutional Networks
• Semi-Supervised Segmentation of Radiation-Induced Pulmonary Fibrosis from Lung CT Scans with Multi-Scale Guided Dense Attention
• Scalable multimodal convolutional networks for brain tumour segmentation
• Evaluation of Multi-Slice Inputs to Convolutional Neural Networks for Medical Image Segmentation
• Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets
• MRI Super-Resolution with GAN and 3D Multi-Level DenseNet: Smaller, Faster, and Better
• Uncertainty in multitask learning: joint representations for probabilistic MR-only radiotherapy planning
• Automatic segmentation method of pelvic floor levator hiatus in ultrasound using a self-normalising neural network
• SKA Science Data Challenge 2: analysis and results
• Improved MR to CT synthesis for PET/MR attenuation correction using Imitation Learning
• Multi-level Context Gating of Embedded Collective Knowledge for Medical Image Segmentation
• m2caiSeg: Semantic Segmentation of Laparoscopic Images using Convolutional Neural Networks
• Whole Brain Segmentation with Full Volume Neural Network
• Simulation of Brain Resection for Cavity Segmentation Using Self-Supervised and Semi-Supervised Learning
• Generalized Wasserstein Dice Loss, Test-time Augmentation, and Transformers for the BraTS 2021 challenge
• Automatic cerebral hemisphere segmentation in rat MRI with lesions via attention-based convolutional neural networks
• 3D Whole Brain Segmentation using Spatially Localized Atlas Network Tiles
• DINs: Deep Interactive Networks for Neurofibroma Segmentation in Neurofibromatosis Type 1 on Whole-Body MRI
• Towards safe deep learning: accurately quantifying biomarker uncertainty in neural network predictions
• Robustness study of noisy annotation in deep learning based medical image segmentation
• Spatially Localized Atlas Network Tiles Enables 3D Whole Brain Segmentation from Limited Data
• A Comprehensive Study on Medical Image Segmentation using Deep Neural Networks
• Robust Automatic Whole Brain Extraction on Magnetic Resonance Imaging of Brain Tumor Patients using Dense-Vnet
• Deep learning-based attenuation correction in the image domain for myocardial perfusion SPECT imaging
• Densely Connected Recurrent Residual (Dense R2UNet) Convolutional Neural Network for Segmentation of Lung CT Images
• Permutohedral Attention Module for Efficient Non-Local Neural Networks
• Distributed Weight Consolidation: A Brain Segmentation Case Study
• Recurrent U-net for automatic pelvic floor muscle segmentation on 3D ultrasound
• Knowing what you know in brain segmentation using Bayesian deep neural networks
• Eisen: a python package for solid deep learning
• Label merge-and-split: A graph-colouring approach for memory-efficient brain parcellation
• Reproducibility Evaluation of SLANT Whole Brain Segmentation Across Clinical Magnetic Resonance Imaging Protocols
• Feature-Fused Context-Encoding Network for Neuroanatomy Segmentation
• ACEnet: Anatomical Context-Encoding Network for Neuroanatomy Segmentation
• Pulse Sequence Resilient Fast Brain Segmentation
• Federated Whole Prostate Segmentation in MRI with Personalized Neural Architectures
• Holistic Decomposition Convolution for Effective Semantic Segmentation of 3D MR Images
• Frequency-Supervised MR-to-CT Image Synthesis
• Beyond one-hot encoding? Journey into compact encoding for large multi-class segmentation
• Enhanced generative adversarial network for 3D brain MRI super-resolution
• Learning joint lesion and tissue segmentation from task-specific hetero-modal datasets
• Deep Boosted Regression for MR to CT Synthesis
• Enhancing Fiber Orientation Distributions using convolutional Neural Networks
• Atlas-aware ConvNetfor Accurate yet Robust Anatomical Segmentation
• CompareNet: Anatomical Segmentation Network with Deep Non-local Label Fusion
• Longitudinal analysis of fetal MRI in patients with prenatal spina bifida repair
• Iteratively-Refined Interactive 3D Medical Image Segmentation with Multi-Agent Reinforcement Learning