MINR: Efficient Implicit Neural Representations for Multi-Image Encoding

TL;DR

MINR introduces a layer-sharing approach for implicit neural representations to efficiently encode multiple images, significantly reducing parameters while maintaining high reconstruction quality and scalability.

Contribution

The paper proposes a novel layer-sharing method for INRs that enables efficient multi-image encoding with reduced parameters and added a projection layer for individual image features.

Findings

Saves up to 60% parameters compared to traditional INRs.

Maintains comparable image reconstruction and super-resolution performance.

Scales effectively to encode 100 images with high PSNR.

Abstract

Implicit Neural Representations (INRs) aim to parameterize discrete signals through implicit continuous functions. However, formulating each image with a separate neural network~(typically, a Multi-Layer Perceptron (MLP)) leads to computational and storage inefficiencies when encoding multi-images. To address this issue, we propose MINR, sharing specific layers to encode multi-image efficiently. We first compare the layer-wise weight distributions for several trained INRs and find that corresponding intermediate layers follow highly similar distribution patterns. Motivated by this, we share these intermediate layers across multiple images while preserving the input and output layers as input-specific. In addition, we design an extra novel projection layer for each image to capture its unique features. Experimental results on image reconstruction and super-resolution tasks demonstrate that MINR can save up to 60\% parameters while maintaining comparable performance. Particularly, MINR scales effectively to handle 100 images, maintaining an average peak signal-to-noise ratio (PSNR) of 34 dB. Further analysis of various backbones proves the robustness of the proposed MINR.