Frame Reconstruction with Missing Data from Multiple Images

To solve the missing data problem that is caused by reasons, such as occlusion, frame reconstruction by a two-level strategy in multiple images was considered. The method first performed a projective reconstruction combining singular value decomposition (SVD) and subspace method with missing data, which estimated projective shape, projection matrices, projective depths and missing data iteratively. Then it converted the projective solution to a Euclidean one with the unknown focal length and the constant principal point by enforcing constraints. Using the constraints and the fact that scale measurement matrix can recover numberless projection matrices and point matrices, the set equations of the transformation matrix from the projective reconstruction to Euclidean reconstruction were obtained. Experimental results using real images are provided to illustrate the performance of the proposed method.