Content-Aware Frequency Encoding for Implicit Neural Representations with Fourier-Chebyshev Features

TL;DR

This paper introduces CAFE, a novel content-aware frequency encoding method for implicit neural representations that adaptively synthesizes and selects frequency bases, improving the capture of high-frequency details and overall performance.

Contribution

We propose CAFE, a flexible frequency encoding framework that combines Fourier and Chebyshev features with learned weights, enhancing INRs' ability to represent complex signals.

Findings

Outperforms existing Fourier-based methods on multiple benchmarks.

Efficiently synthesizes a broad range of frequency bases.

Improves high-frequency detail capture and stability.

Abstract

Implicit Neural Representations (INRs) have emerged as a powerful paradigm for various signal processing tasks, but their inherent spectral bias limits the ability to capture high-frequency details. Existing methods partially mitigate this issue by using Fourier-based features, which usually rely on fixed frequency bases. This forces multi-layer perceptrons (MLPs) to inefficiently compose the required frequencies, thereby constraining their representational capacity. To address this limitation, we propose Content-Aware Frequency Encoding (CAFE), which builds upon Fourier features through multiple parallel linear layers combined via a Hadamard product. CAFE can explicitly and efficiently synthesize a broader range of frequency bases, while the learned weights enable the selection of task-relevant frequencies. Furthermore, we extend this framework to CAFE+, which incorporates Chebyshev features as a complementary component to Fourier bases. This combination provides a stronger and more stable frequency representation. Extensive experiments across multiple benchmarks validate the effectiveness and efficiency of our approach, consistently achieving superior performance over existing methods. Our code is available at https://github.com/JunboKe0619/CAFE.