Search papers, labs, and topics across Lattice.
The paper introduces the Subtractive Modulative Network (SMN), a novel implicit neural representation architecture inspired by subtractive synthesis, designed for parameter-efficient signal representation. SMN employs a learnable periodic activation layer (Oscillator) to create a multi-frequency basis and modulative mask modules (Filters) to generate high-order harmonics. The proposed SMN achieves state-of-the-art performance on image reconstruction (PSNR $40+$ dB) and novel view synthesis tasks while maintaining parameter efficiency.
Achieve state-of-the-art image and NeRF reconstruction with a parameter-efficient implicit neural representation inspired by classical subtractive synthesis.
We propose the Subtractive Modulative Network (SMN), a novel, parameter-efficient Implicit Neural Representation (INR) architecture inspired by classical subtractive synthesis. The SMN is designed as a principled signal processing pipeline, featuring a learnable periodic activation layer (Oscillator) that generates a multi-frequency basis, and a series of modulative mask modules (Filters) that actively generate high-order harmonics. We provide both theoretical analysis and empirical validation for our design. Our SMN achieves a PSNR of $40+$ dB on two image datasets, comparing favorably against state-of-the-art methods in terms of both reconstruction accuracy and parameter efficiency. Furthermore, consistent advantage is observed on the challenging 3D NeRF novel view synthesis task. Supplementary materials are available at https://inrainbws.github.io/smn/.
