Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Category-Fragment Segmentation Framework for Pelvic Fracture Segmentation in X-ray Images

Daiqi Liu, Fuxin Fan, Andreas Maier

TL;DR

The paper tackles automatic segmentation of pelvic fracture fragments in 2D X-ray images, a challenging problem due to projection overlap. It introduces the Category-Fragment Segmentation (CFS) framework, comprising category segmentation via Swin UNETR, fragment segmentation via Mask R-CNN using category masks as inputs, and a post-processing step that multiplies category masks with fragment masks to refine outputs. Evaluated on a DeepDRR-generated MICCAI PengWin dataset (50,000 X-ray images from 100 CT volumes, 448×448 resolution), the approach achieves state-of-the-art-like performance with $IoU=0.914$ for category segmentation and $IoU=0.775$ for fragment segmentation using the best Swin UNETR +p model. The findings highlight the value of region-aware, multi-stage segmentation and post-processing for improving intraoperative pelvic fracture guidance, while also identifying challenges posed by overlapping anatomy and variability in fragment boundaries.

Abstract

Pelvic fractures, often caused by high-impact trauma, frequently require surgical intervention. Imaging techniques such as CT and 2D X-ray imaging are used to transfer the surgical plan to the operating room through image registration, enabling quick intraoperative adjustments. Specifically, segmenting pelvic fractures from 2D X-ray imaging can assist in accurately positioning bone fragments and guiding the placement of screws or metal plates. In this study, we propose a novel deep learning-based category and fragment segmentation (CFS) framework for the automatic segmentation of pelvic bone fragments in 2D X-ray images. The framework consists of three consecutive steps: category segmentation, fragment segmentation, and post-processing. Our best model achieves an IoU of 0.91 for anatomical structures and 0.78 for fracture segmentation. Results demonstrate that the CFS framework is effective and accurate.

A Category-Fragment Segmentation Framework for Pelvic Fracture Segmentation in X-ray Images

TL;DR

The paper tackles automatic segmentation of pelvic fracture fragments in 2D X-ray images, a challenging problem due to projection overlap. It introduces the Category-Fragment Segmentation (CFS) framework, comprising category segmentation via Swin UNETR, fragment segmentation via Mask R-CNN using category masks as inputs, and a post-processing step that multiplies category masks with fragment masks to refine outputs. Evaluated on a DeepDRR-generated MICCAI PengWin dataset (50,000 X-ray images from 100 CT volumes, 448×448 resolution), the approach achieves state-of-the-art-like performance with for category segmentation and for fragment segmentation using the best Swin UNETR +p model. The findings highlight the value of region-aware, multi-stage segmentation and post-processing for improving intraoperative pelvic fracture guidance, while also identifying challenges posed by overlapping anatomy and variability in fragment boundaries.

Abstract

Pelvic fractures, often caused by high-impact trauma, frequently require surgical intervention. Imaging techniques such as CT and 2D X-ray imaging are used to transfer the surgical plan to the operating room through image registration, enabling quick intraoperative adjustments. Specifically, segmenting pelvic fractures from 2D X-ray imaging can assist in accurately positioning bone fragments and guiding the placement of screws or metal plates. In this study, we propose a novel deep learning-based category and fragment segmentation (CFS) framework for the automatic segmentation of pelvic bone fragments in 2D X-ray images. The framework consists of three consecutive steps: category segmentation, fragment segmentation, and post-processing. Our best model achieves an IoU of 0.91 for anatomical structures and 0.78 for fracture segmentation. Results demonstrate that the CFS framework is effective and accurate.

Paper Structure

This paper contains 4 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Materials and methods
  3. Results
  4. Discussion

Figures (2)

  • Figure 1: Overview of the proposed pelvic fracture segmentation framework. The framework is divided into three main components: category segmentation, fragment segmentation, and post-processing.
  • Figure 2: Relationship between anatomical overlap and segmentation performance. blue represents the overlapping regions, red indicates false positives, and green indicates false negatives.