Drexel University’s biomedical engineers have developed a method for reconstructing accurate 3-dimentional surface representations of the ribs from anterior-posterior 2-dimentional radiographs using combined 2D active appearance models and 3D point distribution models. This method was implemented in the MATLAB code to perform the 3D reconstructions with minimal user input and the speed of less than five minutes per rib pair. 

Patient-specific computational models and surgical inspection of the thorax require rendering and visualization of the 3D anatomy. While computer tomography CT-based 3D surface reconstruction is the current clinical standard, it is associated with health risks related to significant radiation exposure. The ability to reconstruct 3D surface anatomy from 2D radiographic images would significantly reduce the risk of radiation exposure and also reduce the time spent in medical facilities. Current 2D radiographs to 3D rib reconstruction methods reported in the literature produce only 3D rib centerlines and do not represent the anatomical rib surface. Our approach is a significant improvement over the existing methods used to render 3D rib surface geometry from 2D radiographs and could serve as the state of the art for future developments in medical image processing.