Abstract
We present a novel 3D ear shape matching algorithm. Firstly, we compute the saliency value of every point in 3D ear point cloud using the Gaussian-weighted average of the mean curvature, sort the points in accordance with their saliency value, Then we get the salient key points set using the poisson sampling. Finally, we fit the neighborhood of each salient key point using quadratic principal manifold, to construct the local shape feature. The experiment results show that, our method has a higher approximation precision and a higher matching accuracy.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have