FLAIR: Frequency- and Locality-Aware Implicit Neural Representations
This addresses a bottleneck in INR methods for vision tasks, offering improved performance but appearing incremental as it builds on the existing INR paradigm.
The paper tackles the problem of Implicit Neural Representations (INRs) lacking frequency selectivity and spatial localization, which causes spectral bias and difficulty capturing high-frequency details, by proposing FLAIR with RC-GAUSS activation and Wavelet-Energy-Guided Encoding, achieving consistent outperformance over existing INRs in 2D image representation/restoration and 3D reconstruction.
Implicit Neural Representations (INRs) leverage neural networks to map coordinates to corresponding signals, enabling continuous and compact representations. This paradigm has driven significant advances in various vision tasks. However, existing INRs lack frequency selectivity, spatial localization, and sparse representations, leading to an over-reliance on redundant signal components. Consequently, they exhibit spectral bias, tending to learn low-frequency components early while struggling to capture fine high-frequency details. To address these issues, we propose FLAIR (Frequency- and Locality-Aware Implicit Neural Representations), which incorporates two key innovations. The first is RC-GAUSS, a novel activation designed for explicit frequency selection and spatial localization under the constraints of the time-frequency uncertainty principle (TFUP). The second is Wavelet-Energy-Guided Encoding (WEGE), which leverages the discrete wavelet transform (DWT) to compute energy scores and explicitly guide frequency information to the network. Our method consistently outperforms existing INRs in 2D image representation and restoration, as well as 3D reconstruction.