Imaging is a keystone for the understanding and delivery of craniofacial health care and recent developments have led to many diverse technologies and approaches. This paper reviews new developments in three-dimensional imaging, as well as three-dimensional facial image acquisition. Visualization and convergence of the data from these technologies are also described for construction of patient-specific models.
We present an algorithm to reduce the number of slices from 2D contour cross sections. The main aim of the algorithm is to filter less significant slices while preserving an acceptable level of output quality and keeping the computational cost to reconstruct surface(s) at a minimal level. This research is motivated mainly by two factors; first 2D cross sections data is often huge in size and high in precisions – the computational cost to reconstruct surface(s) from them is closely related to the size and complexity of this data. Second, we can trades visual fidelity with speed of computations if we can remove visually insignificant data from the original dataset which may contains redundant information. In our algorithm we use the number of contour points on a pair of slices to calculate the distance between them. Selection to retain/reject a slice is based on the value of distance compared against a threshold value. Optimal threshold value is derived to produce set of slices that collectively represent the feature of the dataset. We tested our algorithm over six different set of data, varying in complexities and sizes. The results show slice reduction rate depends on the complexity of the dataset, where highest reduction percentage is achieved for objects with lots of constant local variations. Our derived optimal thresholds seem to be able to produce the right set of slices with the potential of creating surface(s) that traded off the accuracy and speed requirements.
This paper describes the use of stereophotogrammetry approach to measure and hence identify accurately threedimensional (3D) coordinates of important landmarks on a craniofacial surface. A "novel" technique dubbed as "natural features" technique was employed to accurately compute the 3D coordinates of selected craniofacial landmarks. The natural features technique involves the use of 3D coordinates of the natural features (such as acne, scar, corners of eyes, edge of mouth, point of chin, etc.) that appear on the craniofacial surface as an absolute stereophotogrammetric mapping control points. The 3D coordinates of the natural features were gained using digital photogrammetric bundle adjustment method. Validation of the proposed technique has firstly been carried out using mannequin and finally, it was applied on the real-life human faces. The result shows that the craniofacial landmark measurement accuracy of 0.8mm with one standard deviation can be successfully achieved by the proposed technique.