BID-NeRF: RGB-D image pose estimation with inverted Neural Radiance Fields
This work provides incremental improvements to RGB-D image pose estimation for computer vision applications.
The paper tackles the problem of improving pose estimation accuracy and convergence in Inverted Neural Radiance Fields (iNeRF) by introducing depth-based and multi-image loss functions, omitting hierarchical sampling, and extending sampling intervals, resulting in significantly improved convergence speed and substantially extended basin of convergence.
We aim to improve the Inverted Neural Radiance Fields (iNeRF) algorithm which defines the image pose estimation problem as a NeRF based iterative linear optimization. NeRFs are novel neural space representation models that can synthesize photorealistic novel views of real-world scenes or objects. Our contributions are as follows: we extend the localization optimization objective with a depth-based loss function, we introduce a multi-image based loss function where a sequence of images with known relative poses are used without increasing the computational complexity, we omit hierarchical sampling during volumetric rendering, meaning only the coarse model is used for pose estimation, and we how that by extending the sampling interval convergence can be achieved even or higher initial pose estimate errors. With the proposed modifications the convergence speed is significantly improved, and the basin of convergence is substantially extended.