Automatic Liver Lesion Segmentation Using A Deep Convolutional Neural Network Method

Abstract

Liver lesion segmentation is an important step for liver cancer diagnosis, treatment planning and treatment evaluation. LiTS (Liver Tumor Segmentation Challenge) provides a common testbed for comparing different automatic liver lesion segmentation methods. We participate in this challenge by developing a deep convolutional neural network (DCNN) method. The particular DCNN model works in 2.5D in that it takes a stack of adjacent slices as input and produces the segmentation map corresponding to the center slice. The model has 32 layers in total and makes use of both long range concatenation connections of U-Net [1] and short-range residual connections from ResNet [2]. The model was trained using the 130 LiTS training datasets and achieved an average Dice score of 0.67 when evaluated on the 70 test CT scans, which ranked first for the LiTS challenge at the time of the ISBI 2017 conference.

Submission for ISBI'2017 LiTS Challenge ISIC2017

References in corpus (1)

• Improving Computer-aided Detection using Convolutional Neural Networks and Random View Aggregation

Cited by in corpus (26)

• An overview of deep learning in medical imaging focusing on MRI
• ResViT: Residual vision transformers for multi-modal medical image synthesis
• Dose evaluation of fast synthetic-CT generation using a generative adversarial network for general pelvis MR-only radiotherapy
• SD-CNN: a Shallow-Deep CNN for Improved Breast Cancer Diagnosis
• Deep learning-based synthetic-CT generation in radiotherapy and PET: a review
• MRI-only brain radiotherapy: assessing the dosimetric accuracy of synthetic CT images generated using a deep learning approach
• Transfer Learning in Magnetic Resonance Brain Imaging: a Systematic Review
• Attenuation correction for brain PET imaging using deep neural network based on dixon and ZTE MR images
• Self-Attentive Spatial Adaptive Normalization for Cross-Modality Domain Adaptation
• Male pelvic synthetic CT generation from T1-weighted MRI using 2D and 3D convolutional neural networks
• CBCT-to-CT synthesis with a single neural network for head-and-neck, lung and breast cancer adaptive radiotherapy
• DC-cycleGAN: Bidirectional CT-to-MR Synthesis from Unpaired Data
• Generation of abdominal synthetic CTs from 0.35T MR images using generative adversarial networks for MR-only liver radiotherapy
• Classical and learned MR to pseudo-CT mappings for accurate transcranial ultrasound simulation
• Known Operator Learning and Hybrid Machine Learning in Medical Imaging -- A Review of the Past, the Present, and the Future
• Deep Learning based Multi-modal Computing with Feature Disentanglement for MRI Image Synthesis
• Domain Knowledge Driven 3D Dose Prediction Using Moment-Based Loss Function
• Synthesizing CT from Ultrashort Echo-Time MR Images via Convolutional Neural Networks
• Synthetic Elastography using B-mode Ultrasound through a Deep Fully-Convolutional Neural Network
• An Attentive-based Generative Model for Medical Image Synthesis
• MR-Based PET Attenuation Correction using a Combined Ultrashort Echo Time/Multi-Echo Dixon Acquisition
• MRI-based Material Mass Density and Relative Stopping Power Estimation via Deep Learning for Proton Therapy
• HAGAN: Hybrid Augmented Generative Adversarial Network for Medical Image Synthesis
• View it like a radiologist: Shifted windows for deep learning augmentation of CT images
• A Unified Generation-Registration Framework for Improved MR-based CT Synthesis in Proton Therapy
• Accurate 2D Reconstruction for PET Scanners based on the Analytical White Image Model