Robust sparse bounding sphere for 3D face recognition

► A robust sparse bounding sphere representation (RSBSR) is proposed to analyze 3D facial data. ► 3D facial models are projected on the bounding spheres to describe both depth and the 3D geometric shape. ► A robust group sparse regression model (RGSRM) learns the regression matrix. ► Under the constraints of Spectral Regression and corruptions, noise corruptions can be removed.