CBARF: Cascaded Bundle-Adjusting Neural Radiance Fields from Imperfect Camera Poses
This addresses a key limitation in volumetric neural rendering for applications like 3D reconstruction and virtual reality, though it is incremental as it builds on existing NeRF methods.
The paper tackles the problem of synthesizing high-quality novel views from imperfect camera poses in Neural Radiance Fields (NeRF) by proposing CBARF, a framework that optimizes camera poses in a coarse-to-fine manner and reconstructs scenes, achieving state-of-the-art performance in pose optimization and novel view synthesis, especially with large camera pose noise.
Existing volumetric neural rendering techniques, such as Neural Radiance Fields (NeRF), face limitations in synthesizing high-quality novel views when the camera poses of input images are imperfect. To address this issue, we propose a novel 3D reconstruction framework that enables simultaneous optimization of camera poses, dubbed CBARF (Cascaded Bundle-Adjusting NeRF).In a nutshell, our framework optimizes camera poses in a coarse-to-fine manner and then reconstructs scenes based on the rectified poses. It is observed that the initialization of camera poses has a significant impact on the performance of bundle-adjustment (BA). Therefore, we cascade multiple BA modules at different scales to progressively improve the camera poses. Meanwhile, we develop a neighbor-replacement strategy to further optimize the results of BA in each stage. In this step, we introduce a novel criterion to effectively identify poorly estimated camera poses. Then we replace them with the poses of neighboring cameras, thus further eliminating the impact of inaccurate camera poses. Once camera poses have been optimized, we employ a density voxel grid to generate high-quality 3D reconstructed scenes and images in novel views. Experimental results demonstrate that our CBARF model achieves state-of-the-art performance in both pose optimization and novel view synthesis, especially in the existence of large camera pose noise.