Author :
Tang, Li ; Wu, Chengke ; Tsui, H.T. ; Liu, Shigang
Abstract :
A novel algorithm for 3D reconstruction from image sequences is presented. Global optimisation is applied to a rough model by minimising the back projection errors in each image. Different weights are added to distinguish visible and occluded points. Satisfactory results on both simulated data and real images are reported
Keywords :
computer vision <3D reconstruction, both vis. and missing data, algm.>; image reconstruction <3D reconstruction, both vis. and missing data, algm.>; image sequences <3D reconstruction, both vis. and missing data, algm.>; iterative methods <3D reconstruction, both vis. and missing data, algm.>; optimisation <3D reconstruction, both vis. and missing data, algm.>; 3D reconstruction algorithm; back projection errors; computer vision; global optimisation; image sequences; iteration; missing data; occluded points; partial structures; real images; rough 3D model; virtual cameras; visible points; weighting matrix;
