Neural Fourier Filter Bank
This work addresses the need for more efficient and detailed reconstructions in computer vision and graphics, representing an incremental improvement over existing grid-based methods.
The paper tackles the problem of efficient and detailed signal reconstruction by introducing a neural field method that decomposes signals spatially and frequency-wise, outperforming state-of-the-art methods in model compactness and convergence speed on tasks like 2D image fitting, 3D shape reconstruction, and neural radiance fields.
We present a novel method to provide efficient and highly detailed reconstructions. Inspired by wavelets, we learn a neural field that decompose the signal both spatially and frequency-wise. We follow the recent grid-based paradigm for spatial decomposition, but unlike existing work, encourage specific frequencies to be stored in each grid via Fourier features encodings. 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. 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. Our code is available at https://github.com/ubc-vision/NFFB.