Search papers, labs, and topics across Lattice.
This study revisits initialization-free bundle adjustment (InitFree BA) by establishing a unified evaluation framework that combines existing Object-Space Error (OSE) formulations with a controlled dataset generator. The authors reveal a significant optimization-reconstruction gap, demonstrating that low OSE values do not guarantee accurate 3D reconstructions, which is critical for understanding the limitations of current methods. Key factors influencing reconstruction success include initialization priors, landmark observation density, and the stability of metric upgrades, suggesting that minimizing OSE alone is insufficient for reliable outcomes.
Low Object-Space Error doesn't guarantee accurate 3D reconstructions, exposing a critical gap in initialization-free bundle adjustment methods.
Initialization-free bundle adjustment (InitFree BA) aims to recover camera poses and scene structure directly from image observations, avoiding the geometric initialization stages of conventional structure-from-motion pipelines. Recent methods based on Object-Space Error (OSE) formulations and Variable Projection (VarPro) show encouraging optimization behavior from random camera configurations. However, existing evaluations primarily measure optimization success, leaving unclear whether a low OSE objective yields a valid metric 3D reconstruction. We revisit InitFree BA experimentally through a unified evaluation framework combining a C++ implementation of existing OSE formulations with a Blender-based dataset generator providing exact ground truth and controlled camera configurations and observation densities. Our experiments reveal a previously overlooked optimization--reconstruction gap: projective solutions with similarly low OSE values can lead to substantially different Euclidean reconstructions after metric upgrade. We identify initialization priors, landmark observation density, and metric-upgrade stability as key factors governing reconstruction success. Overall, our results suggest that the main challenge of InitFree BA is not merely minimizing OSE objectives, but obtaining projective reconstructions that admit reliable metric upgrade. We believe that the proposed benchmark, implementation, and analysis establish stronger experimental foundations for future research on initialization-free bundle adjustment, a problem largely unexplored within the computer vision community. Project page is available at https://github.com/simonwebertum/InitFreeBA.git.
