CT Image Synthesis Using Weakly Supervised Segmentation and Geometric Inter-Label Relations For COVID Image Analysis

Abstract

While medical image segmentation is an important task for computer aided diagnosis, the high expertise requirement for pixelwise manual annotations makes it a challenging and time consuming task. Since conventional data augmentations do not fully represent the underlying distribution of the training set, the trained models have varying performance when tested on images captured from different sources. Most prior work on image synthesis for data augmentation ignore the interleaved geometric relationship between different anatomical labels. We propose improvements over previous GAN-based medical image synthesis methods by learning the relationship between different anatomical labels. We use a weakly supervised segmentation method to obtain pixel level semantic label map of images which is used learn the intrinsic relationship of geometry and shape across semantic labels. Latent space variable sampling results in diverse generated images from a base image and improves robustness. We use the synthetic images from our method to train networks for segmenting COVID-19 infected areas from lung CT images. The proposed method outperforms state-of-the-art segmentation methods on a public dataset. Ablation studies also demonstrate benefits of integrating geometry and diversity.

arXiv commentarXiv admin note: substantial text overlap with arXiv:2003.14119; text overlap with arXiv:1908.10555, arXiv:2004.14133 by other authors

References in corpus (20)

• Adam: A Method for Stochastic Optimization
• Generative Adversarial Network in Medical Imaging: A Review
• Data Augmentation Generative Adversarial Networks
• COVID-CT-Dataset: A CT Scan Dataset about COVID-19
• Harmony-Search and Otsu based System for Coronavirus Disease (COVID-19) Detection using Lung CT Scan Images
• Synergic Adversarial Label Learning for Grading Retinal Diseases via Knowledge Distillation and Multi-task Learning
• Retinal Vasculature Segmentation Using Local Saliency Maps and Generative Adversarial Networks For Image Super Resolution
• Learning in an Uncertain World: Representing Ambiguity Through Multiple Hypotheses
• Night Time Haze and Glow Removal using Deep Dilated Convolutional Network
• Chest X-rays Classification: A Multi-Label and Fine-Grained Problem
• Progressive Generative Adversarial Networks for Medical Image Super resolution
• GAN Based Medical Image Registration
• Pathological Retinal Region Segmentation From OCT Images Using Geometric Relation Based Augmentation
• Improving Medical Image Classification with Label Noise Using Dual-uncertainty Estimation
• Registration of Histopathogy Images Using Structural Information From Fine Grained Feature Maps
• sZoom: A Framework for Automatic Zoom into High Resolution Surveillance Videos
• Consensus Based Medical Image Segmentation Using Semi-Supervised Learning And Graph Cuts
• Anomaly Detection on X-Rays Using Self-Supervised Aggregation Learning
• Multi-scale Deep Learning Architecture for Nucleus Detection in Renal Cell Carcinoma Microscopy Image
• Relational Subsets Knowledge Distillation for Long-tailed Retinal Diseases Recognition

Cited by in corpus (2)

• Learning of Inter-Label Geometric Relationships Using Self-Supervised Learning: Application To Gleason Grade Segmentation
• Multimodal Generalized Zero Shot Learning for Gleason Grading using Self-Supervised Learning