Segmentation and Fracture Detection in X-ray images for Traumatic Pelvic Injury

Due to the risk of complications such as hemorrhage, severe pelvic trauma is associated with a high mortality rate. Prompt medical treatment is therefore vital. However, the complexity of the injuries can make successful diagnosis and treatment challenging. By generating predictions and recommendations based on patient data, computer-aided decision support systems have the potential to assist physicians in improving outcomes. However, no current system considers features automatically extracted from medical images. This dissertation describes a system to extract diagnostic features from pelvic X-ray images that can be used as input to the prediction process; specifically, the presence of fracture and quantitative measures of displacement. Feature extraction requires prior identification of separate structures of interest within the pelvis. The proposed system therefore incorporates a hierarchical segmentation algorithm which is able to automatically extract multiple structures in a single pass, using a combination of anatomical knowledge and computational techniques such as directed Hough Transform. This algorithm also applies a novel Spline/ASM segmentation method which combines cubic spline interpolation with a deformable model approach which maintains curved contours and provides local control over segmentation. In order for the proposed system to be used as a component in a computerized decision support system, segmentation is designed to be entirely automatic. Furthermore, Spline/ASM is suitable for many other segmentation applications where the objects of interest show curved contours. After successful segmentation, fracture detection is performed on the pelvic ring and pubis structures, using an algorithm based on wavelet transform, anatomical information and boundary tracing. A method is also developed to calculate quantitative measures of symphysis pubis displacement that may indicate pelvic instability and prove useful in identifying fracture patterns. Finally, X-ray features are combined with patient demographics and physiological scores for generation of predictive rules for injury severity, with promising current results. This indicates the potential diagnostic value of the extracted features, and in turn the usefulness of the proposed radiograph analysis component in a larger decision support system.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science

Year dc:date.available

2010

Author dc:creator

• Smith, Rebecca

Contributors dc:contributor

• Kayvan Najarian

• image processing
• x-ray analysis
• computer-aided diagnosis
• machine learning
• image analysis
• Computer Sciences
• Physical Sciences and Mathematics

Statement dc:rights

• © The Author

Identifier

https://scholarscompass.vcu.edu/etd/76

OAI identifier oai:identifier

oai:scholarscompass.vcu.edu:etd-1075