Rεςεαρch

Automatic detection of direct radiation for digital fluoroscopy optimization

In real-time dynamic X-ray examination, the technician moves the patient around to get a desirable view of the anatomy in motion while the beam is activated. As a result, the static region of interest (ROI) may capture directly radiated areas on the detector, yielding overshot statistic estimates. We tackle this problem by implementing a histogram-based real-time solution. In the transformed domain a histogram is computed, from which discriminant features are extracted. Then each image is classified as containing partial, all, or null direct radiation. According to the image characteristics, the appropriate threshold value is set. Our direct radiation detection (DRD) method leverages the power of model matching, machine learning and domain knowledge to characterize and segment images using histograms. To the best of our knowledge, our work is the first to address the direct radiation detection problem.