# CBARF: Cascaded Bundle-Adjusting Neural Radiance Fields from Imperfect Camera Poses

arXiv: https://arxiv.org/abs/2310.09776 · HTML: https://mlsanity.com/paper/2310.09776/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 32 of its 87 references that are in the map.
- First posted on arXiv 15 Oct 2023.

## Builds on (most cited first, 30 of 32)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Occupancy Networks: Learning 3D Reconstruction in Function Space](https://mlsanity.com/paper/1812.03828/) (2018)
- [Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields](https://mlsanity.com/paper/2103.13415/) (2021)
- [Plenoxels: Radiance Fields without Neural Networks](https://mlsanity.com/paper/2112.05131/) (2021)
- [pixelNeRF: Neural Radiance Fields from One or Few Images](https://mlsanity.com/paper/2012.02190/) (2020)
- [D-NeRF: Neural Radiance Fields for Dynamic Scenes](https://mlsanity.com/paper/2011.13961/) (2020)
- [NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections](https://mlsanity.com/paper/2008.02268/) (2020)
- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction](https://mlsanity.com/paper/2111.11215/) (2021)
- [Multi-Stage Progressive Image Restoration](https://mlsanity.com/paper/2102.02808/) (2021)
- [BARF: Bundle-Adjusting Neural Radiance Fields](https://mlsanity.com/paper/2104.06405/) (2021)
- [Cyclical Learning Rates for Training Neural Networks](https://mlsanity.com/paper/1506.01186/) (2015)
- [3D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstruction](https://mlsanity.com/paper/1604.00449/) (2016)
- [BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networks](https://mlsanity.com/paper/1911.10127/) (2019)
- [Free View Synthesis](https://mlsanity.com/paper/2008.05511/) (2020)
- [SPARF: Neural Radiance Fields from Sparse and Noisy Poses](https://mlsanity.com/paper/2211.11738/) (2022)
- [A Survey of Structure from Motion](https://mlsanity.com/paper/1701.08493/) (2017)
- [NeRF--: Neural Radiance Fields Without Known Camera Parameters](https://mlsanity.com/paper/2102.07064/) (2021)
- [Learning non-maximum suppression](https://mlsanity.com/paper/1705.02950/) (2017)
- [Image-based 3D Object Reconstruction: State-of-the-Art and Trends in the Deep Learning Era](https://mlsanity.com/paper/1906.06543/) (2019)
- [GARF: Gaussian Activated Radiance Fields for High Fidelity Reconstruction and Pose Estimation](https://mlsanity.com/paper/2204.05735/) (2022)
- [DBARF: Deep Bundle-Adjusting Generalizable Neural Radiance Fields](https://mlsanity.com/paper/2303.14478/) (2023)
- [Transformation-Grounded Image Generation Network for Novel 3D View Synthesis](https://mlsanity.com/paper/1703.02921/) (2017)
- [Local-to-Global Registration for Bundle-Adjusting Neural Radiance Fields](https://mlsanity.com/paper/2211.11505/) (2022)
- [CamP: Camera Preconditioning for Neural Radiance Fields](https://mlsanity.com/paper/2308.10902/) (2023)
- [Generalizable Neural Radiance Fields for Novel View Synthesis with Transformer](https://mlsanity.com/paper/2206.05375/) (2022)
- [Hybrid Rotation Averaging: A Fast and Robust Rotation Averaging Approach](https://mlsanity.com/paper/2101.09116/) (2021)
- [SparseFusion: Dynamic Human Avatar Modeling from Sparse RGBD Images](https://mlsanity.com/paper/2006.03630/) (2020)
- [Prior-Guided Multi-View 3D Head Reconstruction](https://mlsanity.com/paper/2107.04277/) (2021)
- [Deep View Synthesis via Self-Consistent Generative Network](https://mlsanity.com/paper/2101.10844/) (2021)

## Built on by (most cited first, 11 of 11)

- [FlowMap: High-Quality Camera Poses, Intrinsics, and Depth via Gradient Descent](https://mlsanity.com/paper/2404.15259/) (2024)
- [TrackNeRF: Bundle Adjusting NeRF from Sparse and Noisy Views via Feature Tracks](https://mlsanity.com/paper/2408.10739/) (2024)
- [ZeroGS: Training 3D Gaussian Splatting from Unposed Images](https://mlsanity.com/paper/2411.15779/) (2024)
- [SceneCompleter: Dense 3D Scene Completion for Generative Novel View Synthesis](https://mlsanity.com/paper/2506.10981/) (2025)
- [URS-NeRF: Unordered Rolling Shutter Bundle Adjustment for Neural Radiance Fields](https://mlsanity.com/paper/2403.10119/) (2024)
- [AttentionGS: Towards Initialization-Free 3D Gaussian Splatting via Structural Attention](https://mlsanity.com/paper/2506.23611/) (2025)
- [Density Uncertainty Quantification with NeRF-Ensembles: Impact of Data and Scene Constraints](https://mlsanity.com/paper/2312.14664/) (2023)
- [PairDropGS: Paired Dropout-Induced Consistency Regularization for Sparse-View Gaussian Splatting](https://mlsanity.com/paper/2605.12072/) (2026)
- [NeRF-based Spacecraft Reconstruction from Monocular Imagery Under Illumination Variability and Pose Uncertainty](https://mlsanity.com/paper/2605.18447/) (2026)
- [Ground4D: Consistency-Aware 4D Reconstruction from Monocular Video](https://mlsanity.com/paper/2606.28828/) (2026)
- [GrainGS: Gradient-Decoupled Gaussian Splatting for Efficient Dynamic Novel View Synthesis](https://mlsanity.com/paper/2607.21448/) (2026)

## Abstract (arXiv)

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.