# DINER: Disorder-Invariant Implicit Neural Representation

Source: [DINER: Disorder-Invariant Implicit Neural Representation](<https://www.sota2.com/research/paper/68fa6de41e209131cea3cb6c>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-01T15:41:25.845Z

Authors: Shaowen Xie, Hao Zhu, Zhen Liu, Qi Zhang, You Zhou, Xun Cao, Zhan Ma

Published: 2022-11-15T03:34:24.000Z

## Abstract

Implicit neural representation (INR) characterizes the attributes of a signal as a function of corresponding coordinates which emerges as a sharp weapon for solving inverse problems. However, the capacity of INR is limited by the spectral bias in the network training. In this paper, we find that such a frequency-related problem could be largely solved by re-arranging the coordinates of the input signal, for which we propose the disorder-invariant implicit neural representation (DINER) by augmenting a hash-table to a traditional INR backbone. Given discrete signals sharing the same histogram of attributes and different arrangement orders, the hash-table could project the coordinates into the same distribution for which the mapped signal can be better modeled using the subsequent INR network, leading to significantly alleviated spectral bias. Experiments not only reveal the generalization of the DINER for different INR backbones (MLP vs. SIREN) and various tasks (image/video representation, phase retrieval, and refractive index recovery) but also show the superiority over the state-of-the-art algorithms both in quality and speed.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [3D Video Representation on UVG ReadySetGo and ShakeNDry](<https://www.sota2.com/research/sota/3d-video-representation-on-uvg-readysetgo-and-shakendry.md>): 3D Video Representation; UVG ReadySetGo and ShakeNDry; PSNR: 50.74.

- [2D image fitting on NTIRE 2017 (train)](<https://www.sota2.com/research/sota/2d-image-fitting-on-ntire-2017-train.md>): 2D Image Fitting; NTIRE 2017 (train); PSNR: 38.67.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2211.07871>)

[PDF](<https://arxiv.org/pdf/2211.07871.pdf>)