HOSC: A Periodic Activation Function for Preserving Sharp Features in Implicit Neural Representations
This is an incremental improvement for researchers and practitioners using implicit neural representations to model signals like images or geometries.
The paper tackles the problem of capturing sharp features in implicit neural representations by introducing the Hyperbolic Oscillation (HOSC) activation function with a controllable sharpness parameter, resulting in empirically improved representation quality in benchmarks.
Recently proposed methods for implicitly representing signals such as images, scenes, or geometries using coordinate-based neural network architectures often do not leverage the choice of activation functions, or do so only to a limited extent. In this paper, we introduce the Hyperbolic Oscillation function (HOSC), a novel activation function with a controllable sharpness parameter. Unlike any previous activations, HOSC has been specifically designed to better capture sudden changes in the input signal, and hence sharp or acute features of the underlying data, as well as smooth low-frequency transitions. Due to its simplicity and modularity, HOSC offers a plug-and-play functionality that can be easily incorporated into any existing method employing a neural network as a way of implicitly representing a signal. We benchmark HOSC against other popular activations in an array of general tasks, empirically showing an improvement in the quality of obtained representations, provide the mathematical motivation behind the efficacy of HOSC, and discuss its limitations.