An example method of generating three-dimensional visual objects representing a person based on two-dimensional images of at least a part of the person's body may include receiving a first polygonal mesh representing a human body part, wherein the first polygonal mesh is compliant with a target application topology. The example method may further include receiving a second polygonal mesh representing the human body part, wherein the second polygonal mesh is derived from a plurality of images of a person. The example method may further include modifying at least one of the first polygonal mesh or the second polygonal mesh to optimize a value of a metric reflecting a difference between the first polygonal mesh and the second polygonal mesh.