3D measurements from X-ray images and dense surface mappings

The combination of multiple conventional X-ray images from a same patient enables the derivation of valuable extra clinical information. When these images are geometrically correlated, common points and regions of interest can be easily identified, contributing to the disclosure of new diagnostic data. We present a very accurate method to estimate the necessary geometrical information with the help of a depth camera and a novel algorithm for surface dense mapping (DSM) of the patient's volume. Test cases with a phantom and a real patient are presented. Results show we can match 2D spots and areas between consecutive X-ray snapshots with a good level of precision. Accurate 3D reconstruction from image pairs is also feasible. The proposed augmented X-ray imaging technique can be an alternative to CT examinations and more expensive and complex modalities like tomosynthesis.