Masoud Ebadi - M.Sc. Student, Department of Electrical and Computer Engineering, Semnan University, Semnan, Iran
Kourosh Kiani - Associate Professor, Department of Electrical and Computer Engineering, Semnan University, Semnan, Iran
Razieh Rastgoo - Assistant Professor, Department of Electrical and Computer Engineering, Semnan University, Semnan, Iran

Mesh saliency has been extensively considered as the measure of visual importance ofcertain parts of ۳D geometries with respect to human visual perception. Previous works of saliencydetection in ۲D images and ۳D images face the challenge of high computational complexity. Totackle this challenge, we propose a simple and fast method for saliency detection using the vertexnormal vector and the parameters of the probability distribution function of their direction change.Due to using the probability distribution of changes instead of computational geometry-basedmodels, we rarely face problems such as sensitivity to scale and neighborhood size in the curvature.Results show that the proposed model obtains comparable results with the previous works forsaliency detection without any additional intermediate calculations. Our method is robust to noise,deformation and triangle count

Mesh, Saliency detection, Normal vector, Simplification, Fast.