Data Transformation Approach Paper & Methodology (aw + pw + awin + pwin + s1 Fac) to FFac GANs [145–151] DM, CycleGAN [152] (aw + pw + awin + pwin + s1 Fac) to RFac pix2pix [153] FFac to FFac CycleGAN (pwin), and component elements (s1 Fac), represented as (aw + (pw) + awin + (pwin) + s1 Fac) to FFac [145–152]. Furthermore, complete facade and roof images for all four directions of a building can be generated Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pages 11956– 11965, 2021. [145] Zhenlong Du, Haiyang Shen, Xiaoli Li, and Meng Wang. 3d building fabrication with geometry and texture