Boosting Segmentation Performance across Datasets using Histogram Specification with Application to Pelvic Bone Segmentation
Authors:
Open Access
- 9 December 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1364-1369
Abstract
Accurate segmentation of the pelvic computerized tomography (CT) imaging is crucial for the clinical diagnosis of pelvic bone diseases and for planning patient-specific hip surgeries. Several methodologies have been proposed for such segmentation tasks with the emergence and advancements of deep learning for digital healthcare. But in a low data scenario, the lack of abundant data needed to train a Deep Neural Network is a significant bottleneck. In this work, we propose a methodology based on modulation of image tonal distributions and deep learning to boost the performance of networks trained on limited data. The strategy involves pre-processing of test data through histogram specification. This simple yet effective approach can be viewed as a style transfer methodology. The segmentation task uses a U-Net configuration with an EfficientNet-B0 backbone, optimized using an augmented BCE-IoU loss function. This configuration is validated on a total of 284 images taken from two publicly available CT datasets, The Cancer Imaging Archive (TCIA) and the Visible Human Project (VHBD). The average performance measures for the Dice coefficient and Intersection over Union are 95.7% and 91.9%, respectively, which give strong evidence for the effectiveness of the approach, which is highly competitive with state-of-the-art methodologies.Keywords
All Related Versions
Funding Information
- Mitacs
This publication has 20 references indexed in Scilit:
- Deep learning to segment pelvic bones: large-scale CT datasets and baseline modelsInternational Journal of Computer Assisted Radiology and Surgery, 2021
- 3D U-net with Multi-level Deep Supervision: Fully Automatic Segmentation of Proximal Femur in 3D MR Images2017
- A Multi-atlas Approach to Region of Interest Detection for Medical Image Classification2017
- MASCG: Multi-Atlas Segmentation Constrained Graph method for accurate segmentation of hip CT imagesMedical Image Analysis, 2015
- The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information RepositoryJournal of Digital Imaging, 2013
- Automated CT Segmentation of Diseased Hip Using Hierarchical and Conditional Statistical Shape Models2013
- Incidence and Total Lifetime Costs of Motor Vehicle–Related Fatal and Nonfatal Injury by Road User Type, United States, 2005Traffic Injury Prevention, 2010
- Performance measure characterization for evaluating neuroimage segmentation algorithmsNeuroImage, 2009
- Coupling Deformable Models for Multi-object SegmentationLecture Notes in Computer Science, 2008
- The Visible Human ProjectProceedings of the IEEE, 1998