A missing data estimation approach for small size image sequence

Data missing is a frequently encountered problem for structure-from-motion (SFM). In this paper, a sub-sequence based approach is proposed to deal with the missing data estimation problem for small size image sequence. In the proposed method, the sub-sequences are first extracted from the original sequence. Further, multiple weaker estimators are constructed by means of the column space fitting (CSF) algorithm. Finally, the missing entries are estimated by a linear programming based weighted model. Experimental results on several widely used image sequences demonstrate the effectiveness and feasibility of the proposed algorithm.

I. Introduction

As one of the most important issues in computer vision, structure from motion (SFM) has been studied extensively in computer vision [1], [2]. For SFM, the feature points may leave the field of view, or may be blocked by other objects in the tracking process [3]. When encountered such a condition, the feature points are invisible, and the corresponding measurements are missing. In the presence of missing data, some typical approaches of SFM, e.g. the factorization algorithm, cannot be used directly. Therefore, missing data estimation is generally a necessary preprocessing step of SFM.