Office of Technology Transfer – University of Michigan

Segmentation and Fracture Detection in Title CT Images for Traumatic Pelvic Injuries

Technology #6008

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Kevin Ward
Managed By
Drew Bennett
Associate Director - Software Licensing 734-615-4004
Patent Protection
US Patent Pending
US Patent Pending

Pelvic injuries—frequently caused by high-energy traumatic events such as a car collision—are considered to be life-threatening due to high risks of damage to major blood vessels and organs in proximity to the pelvis. Injuries associated with pelvic trauma include venous hemorrhage, visceral injury, bladder injury, and nerve deficits. Given its speed, low cost, and minimal disturbance to the patient, X-ray imaging is the first step for diagnosing fractures requiring prompt treatment. However, low resolution and anatomical complexity of the pelvis can make fractures difficult to identify. Computed tomography (CT) scans are an alternative imaging modality for diagnosis of cases with greater complications, including hip fractures and internal bleeding. Given that CT scans capture both cortical and cancellous bone, which have different densities, identification and segmentation of the bone is a challenging task in pelvic fracture detection. An automated system capable of rapid pelvic fracture identification pelvic fractures would be highly relevant in trauma-based situations that require immediate diagnosis.

Segmentation and Fracture Detection in Title CT Images for Traumatic Pelvic Injuries

Accurate detection of pelvic fractures can be performed using automated X-ray and CT images, given that some features may be visible with one imaging modality but not with the other, making the extraction of relevant data from both X-ray and CT images highly useful. Using computational techniques, the X-ray component allows for regions of the pelvis to be segmented and identified, while the CT processing component determines areas of bone and differentiates osseous tissue from non-bone objects. Presence of irregularities automatically identified by the CT component reduces the amount of time needed to perform a pelvic CT scan examination.


  • Detection of pelvic fractures, open/closed internal bleeding for diagnosis and recommended therapy
  • Potential adaptation for use in assessment of fractures in other bone structures


  • Pelvic fracture detection with two different modalities (i.e. X-ray and computed tomography (CT) images)
  • Reduced time needed for full examination of pelvic CT scan
  • Trauma-decision support system utilizes inputs of physiological signals, injury details, demographic information