Quantitative and qualitative comparison of volumetric and surface rendering techniques

This paper compares the fidelity of visualizing craniofacial features using two modern three-dimensional (3D) imaging algorithms emdash one employing surface and the other volume rendering. Each rendering technique was evaluated for its ability to display closed cranial sutures, loss of thin bone through partial volume averaging, and the presence of artifacts. Linear measurements of the orbits, foramina, and mounting holes were taken on the 3D renderings and compared with direct measurements. Both techniques visualized the closed cranial sutures, orbits, mandibles, and teeth. The errors in linear measurement averaged less than 1.5 mm (root mean square) and were not statistically different between the two techniques. Errors are attributable to uncertainty in locating edges due to partial transparency (volume rendering) and sub-optimal lighting. Both rendering techniques suffer from step pattern and thin bone artifacts. The authors conclude that an algorithm for surface construction can provide detailed and accurate representation of the craniofacial anatomy