CoMES: A Novel Method for Robust Nose Tip Detection in Face Range Images

As the most distinct feature point in facial landmarks, nose tip plays a significant role in 3D facial studies. Successful detection of nose tip can facilitate many 3D facial studies tasks. In this paper, we propose a novel method to detect nose tip robustly. The method is robust to noise, need not training, can handle large rotations and occlusions. We first remove small isolated connected regions and noise from the input range image, then establish scale-space by robust smoothing the preprocessed range image. In each scale of the scale-space, we compute multi-angle energy of each point, then we use hierarchical clustering method to cluster the points whose multi-angle energies are larger than a threshold value. In the largest cluster, we can find one point with the largest multi-angle energy. For all scales of the scale-space, we get a series of such points and apply hierarchical clustering again for these points, nose tip will have the largest multi-angle energy in the largest cluster. We evaluate our method in FRGC v2.0 3D face database and BOSPHORUS 3D face database. The experimental results verify the robustness of our method with a high nose tip detection rate.