Abstract: A device for performing multi-pass object rendering using a three-dimensional geometric constraint may include at least one processor configured to receive a mesh of points corresponding to a head of a user. The at least one processor may be further configured to render an image of a sphere and to render elements corresponding to facial features based at least in part on the mesh of points. The at least one processor may be further configured to render an element visibility mask based at least in part on the mesh of points, the element visibility mask being constrained to the surface of the sphere. The at least one processor may be further configured to composite the sphere, the elements, and the element visibility mask to generate an output image. The at least one processor may be further configured to provide the output image for display.