1: \begin{abstract}

2: We present a novel method to provide efficient and highly detailed reconstructions.

3: Inspired by wavelets, we learn a neural field that decompose the signal both spatially and frequency-wise.

4: We follow the recent grid-based paradigm for spatial decomposition, but unlike existing work, 

5: encourage specific frequencies to be stored in each grid via Fourier features encodings.

6: We then apply a multi-layer perceptron with sine activations, taking these Fourier encoded features in at appropriate layers so that higher-frequency components are accumulated on top of lower-frequency components sequentially, which we sum up to form the final output.

7: We demonstrate that our method outperforms the state of the art regarding model compactness and convergence speed on multiple tasks: 2D image fitting, 3D shape reconstruction, and neural radiance fields.

8: Our code is available at 

9: \url{https://github.com/ubc-vision/NFFB}.

10: \end{abstract}

11: