2019/02/26 by Suryansh Kumar, Kumar, Suryansh · 1 citation
Computer Science · Engineering · #Advanced Vision and Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Optical measurement and interference techniques #Robotics and Sensor-Based Localization
paper · pdf · doi:10.48550/arxiv.1902.10274
openalex publication_date 2019/02/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
A simple prior free factorization algorithm citedai2014simple is quite\noften cited work in the field of Non-Rigid Structure from Motion (NRSfM). The\nbenefit of this work lies in its simplicity of implementation, strong\ntheoretical justification to the motion and structure estimation, and its\ninvincible originality. Despite this, the prevailing view is, that it performs\nexceedingly inferior to other methods on several benchmark datasets\n citejensen2018benchmark,akhter2009nonrigid. However, our subtle\ninvestigation provides some empirical statistics which made us think against\nsuch views. The statistical results we obtained supersedes Dai itet\nal. citedai2014simple originally reported results on the benchmark datasets\nby a significant margin under some elementary changes in their core algorithmic\nidea citedai2014simple. Now, these results not only exposes some unrevealed\nareas for research in NRSfM but also give rise to new mathematical challenges\nfor NRSfM researchers. We argue that by \properly utilizing the\nwell-established assumptions about a non-rigidly deforming shape i.e, it\ndeforms smoothly over frames citerabaud2008re and it spans a low-rank space,\nthe simple prior-free idea can provide results which is comparable to the best\navailable algorithms. In this paper, we explore some of the hidden intricacies\nmissed by Dai itet. al. work citedai2014simple and how some elementary\nmeasures and modifications can enhance its performance, as high as approx. 18 %\non the benchmark dataset. The improved performance is justified and empirically\nverified by extensive experiments on several datasets. We believe our work has\nboth practical and theoretical importance for the development of better NRSfM\nalgorithms.\n