Applying Predictive Modeling and Computer Vision (CV) Tools Useful in Biomedicine and Public Health

I will present two examples applying imaging analytics tools to build statistical, predictive models. One example helps diagnose and treat patients with traumatic, penetrating abdominal and pelvic injuries (PAPIs). The other example finds target sites to kill prostate cancer tumors.

The first example describes ways of converting medical image data to spot key signs of PAPIs that help radiologists and clinicians diagnose and decide which patients needed surgery, non-operative management, or observation after trauma-center admission.

The other example used computer vision to output MRI images with fiducial markers that show where to aim external-beam radiation doses to kill tumors, sparing healthy cells. Next, I compute estimates of fiducial marker intensities which establishes tolerance intervals for fiducial marker identification with high degrees of confidence useful in future applications.

2023 Single Track