Methods and apparatus of processing cube face images are disclosed. According one method, each set of six cubic faces is converted into one rectangular assembled image by assembling each set of six cubic faces to maximize a number of continuous boundaries and to minimize a number of discontinuous boundaries. Each continuous boundary corresponds to one boundary between two connected faces with continuous contents from one face to another face. Each discontinuous boundary corresponds to one boundary between two connected faces with discontinuous contents from one face to another face. The method may further comprise applying video coding to the video sequence outputting the compressed data of the video sequence. According to another method, a fully-connected cubic-face image representing an unfolded image from the six faces of the cube is generated and the blank areas are filled with padding data to form a rectangular assembled image.