The author proposes a method for reconstructing 3D human body models from multiple uncalibrated camera views, showcasing superior performance and scalability. The approach involves single-view encoding followed by multi-view fusion, emphasizing the importance of dynamic reweighting networks.
다중 미교정 카메라를 사용한 강인한 3D 인체 복원 방법 소개
Effiziente 3D-Rekonstruktion des menschlichen Körpers aus unkalibrierten Kameramischungen.