Remark on the Successive Projection Algorithm for the Multiple-Sets Split Feasibility Problem