Optimization based fingerprint direction field estimation.

Fingerprint direction field estimation is a key step among most of fingerprint enhancement systems, and so far, in most of systems it heavily relies on local ridge gradient information. However, this method is sensitive to noises and often results in poor estimation in regions where there are too much noises. In this paper, we introduce an effective algorithm to estimate the fingerprint direction field accurately. By using the global ridge characters, the fingerprint direction field is optimized from squared gradients. Experimental results indicate that our algorithm can improve the accuracy of the ridge direction estimation and reduce the influence of noises effectively.

Interactive tooth segmentation of dental models.

The accurate segmentation of the teeth from the triangle mesh is an important step in computer-aided orthodontic. Because teeth come in different shapes and their arrangements vary substantially from one individual to another, tooth segmentation is difficult. This paper proposes a new method to accurately segment the teeth interactively. Based on curvature values of the triangle mesh, feature points are connected to feature regions. After feature lines are extracted from regions, feature contour can be obtained with the help of user supplied information. Using feature contour, the tooth are segmented accurately and individually.