![[2601.15235] Tracing 3D Anatomy in 2D Strokes: A Multi-Stage Projection Driven Approach to Cervical Spine Fracture Identification](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2601.15235] Tracing 3D Anatomy in 2D Strokes: A Multi-Stage Projection Driven Approach to Cervical Spine Fracture Identification
Summary
This article presents a novel approach for identifying cervical spine fractures using a multi-stage projection method that combines 2D imaging with advanced machine learning techniques.
Why It Matters
Cervical spine fractures are critical injuries that require timely diagnosis for effective treatment. This research introduces an innovative method that enhances detection efficiency while reducing computational demands, potentially improving clinical outcomes in radiology.
Key Takeaways
- The proposed method achieves a 3D mIoU of 94.45% using optimized 2D projections.
- The DenseNet121-Unet model provides a Dice score of 87.86% for vertebra segmentation.
- An ensemble approach yields competitive F1 scores of 68.15 and 82.26 for vertebra-level and patient-level evaluations.
- The study includes an explainability component, visualizing key anatomical regions for diagnosis.
- Interobserver analysis shows the model's performance is comparable to expert radiologists.
Computer Science > Computer Vision and Pattern Recognition arXiv:2601.15235 (cs) [Submitted on 21 Jan 2026 (v1), last revised 15 Feb 2026 (this version, v2)] Title:Tracing 3D Anatomy in 2D Strokes: A Multi-Stage Projection Driven Approach to Cervical Spine Fracture Identification Authors:Fabi Nahian Madhurja, Rusab Sarmun, Muhammad E. H. Chowdhury, Adam Mushtak, Israa Al-Hashimi, Sohaib Bassam Zoghoul View a PDF of the paper titled Tracing 3D Anatomy in 2D Strokes: A Multi-Stage Projection Driven Approach to Cervical Spine Fracture Identification, by Fabi Nahian Madhurja and 5 other authors View PDF HTML (experimental) Abstract:Cervical spine fractures are critical medical conditions requiring precise and efficient detection for effective clinical management. This study explores the viability of 2D projection-based vertebra segmentation for vertebra-level fracture detection in 3D CT volumes, presenting an end-to-end pipeline for automated analysis of cervical vertebrae (C1-C7). By approximating a 3D volume through optimized 2D axial, sagittal, and coronal projections, regions of interest are identified using the YOLOv8 model from all views and combined to approximate the 3D cervical spine area, achieving a 3D mIoU of 94.45 percent. This projection-based localization strategy reduces computational complexity compared to traditional 3D segmentation methods while maintaining high performance. It is followed by a DenseNet121-Unet-based multi-label segmentation leveraging varia...