Using X-ray imaging model to improve guidewire detection

Guidewire segmentation is the first step in producing automated tracking system to aid the surgeon during interventions. Most of the currently proposed methods for guidewire detection do not take into account X-ray imaging model and thus perform sub-optimally. We describe a simple imaging model and show how line detection technique based on the eigneanalysis of the Hessian matrix can be adapted to the model. Such adaptation improves detection results and allows application for a wider range of input images. Furthermore, we show how to choose an optimal set of parameters for the detection thus making the method fully automated.