Discriminative Localization in CNNs for Weakly-Supervised Segmentation of Pulmonary Nodules

Abstract

Automated detection and segmentation of pulmonary nodules on lung computed tomography (CT) scans can facilitate early lung cancer diagnosis. Existing supervised approaches for automated nodule segmentation on CT scans require voxel-based annotations for training, which are labor- and time-consuming to obtain. In this work, we propose a weakly-supervised method that generates accurate voxel-level nodule segmentation trained with image-level labels only. By adapting a convolutional neural network (CNN) trained for image classification, our proposed method learns discriminative regions from the activation maps of convolution units at different scales, and identifies the true nodule location with a novel candidate-screening framework. Experimental results on the public LIDC-IDRI dataset demonstrate that, our weakly-supervised nodule segmentation framework achieves competitive performance compared to a fully-supervised CNN-based segmentation method.

References in corpus (3)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: the LUNA16 challenge
• Fully Convolutional Networks for Semantic Segmentation

Cited by in corpus (20)

• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Deep Weakly-Supervised Learning Methods for Classification and Localization in Histology Images: A Survey
• COVID-MTL: Multitask Learning with Shift3D and Random-weighted Loss for Automated Diagnosis and Severity Assessment of COVID-19
• Transformation Consistent Self-ensembling Model for Semi-supervised Medical Image Segmentation
• Attention Based Glaucoma Detection: A Large-scale Database and CNN Model
• Learning Fuzzy Clustering for SPECT/CT Segmentation via Convolutional Neural Networks
• CELNet: Evidence Localization for Pathology Images using Weakly Supervised Learning
• GASNet: Weakly-supervised Framework for COVID-19 Lesion Segmentation
• An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localization
• DeepEM: Deep 3D ConvNets With EM For Weakly Supervised Pulmonary Nodule Detection
• Volumetric Lung Nodule Segmentation using Adaptive ROI with Multi-View Residual Learning
• Estimating brain age based on a healthy population with deep learning and structural MRI
• Interpretation of Disease Evidence for Medical Images Using Adversarial Deformation Fields
• Deep Learning for Automated Medical Image Analysis
• Weakly-Supervised Learning-Based Feature Localization in Confocal Laser Endomicroscopy Glioma Images
• Manifold-driven Attention Maps for Weakly Supervised Segmentation
• Discriminative analysis of the human cortex using spherical CNNs - a study on Alzheimer's disease diagnosis
• When Weak Becomes Strong: Robust Quantification of White Matter Hyperintensities in Brain MRI scans
• Vessel-CAPTCHA: an efficient learning framework for vessel annotation and segmentation
• MED-TEX: Transferring and Explaining Knowledge with Less Data from Pretrained Medical Imaging Models