Ultra-Efficient Reconstruction of Anisotropic Hyperuniform Continuous Random Fields in 2D and 3D via Generalized Spectral Filtering
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911147857281024 |
|---|---|
| author | Zhong, Liyu Mao, Sheng |
| author_facet | Zhong, Liyu Mao, Sheng |
| contents | Hyperuniform continuous random fields suppress large-scale fluctuations while preserving rich local disorder, making them highly attractive for next-generation photonic, thermal and mechanical materials. However, traditional reconstruction techniques often suffer from limited spectral control or excessive computational cost, especially in high-resolution 2D and 3D settings. In this work, we present an ultra-efficient generative algorithm based on generalized superellipse spectral filtering, which allows independent tuning of isotropic and anisotropic spectral envelopes without resorting to costly iterative schemes. We demonstrate our method on a comprehensive set of 2D and 3D examples, showing precise manipulation of spectral band shape and orders-of-magnitude speedup compared to existing approaches. Furthermore, we explore the effect of simple thresholding on the generated fields, analyzing the morphological features and power-spectrum characteristics of the resulting two-phase maps. Our results confirm that the proposed framework not only accelerates hyperuniform field synthesis but also provides a versatile platform for systematic study of binary microstructures derived from continuous designs. This work opens new avenues for large-scale simulation and optimized design of advanced hyperuniform materials. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_08675 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Ultra-Efficient Reconstruction of Anisotropic Hyperuniform Continuous Random Fields in 2D and 3D via Generalized Spectral Filtering Zhong, Liyu Mao, Sheng Materials Science Hyperuniform continuous random fields suppress large-scale fluctuations while preserving rich local disorder, making them highly attractive for next-generation photonic, thermal and mechanical materials. However, traditional reconstruction techniques often suffer from limited spectral control or excessive computational cost, especially in high-resolution 2D and 3D settings. In this work, we present an ultra-efficient generative algorithm based on generalized superellipse spectral filtering, which allows independent tuning of isotropic and anisotropic spectral envelopes without resorting to costly iterative schemes. We demonstrate our method on a comprehensive set of 2D and 3D examples, showing precise manipulation of spectral band shape and orders-of-magnitude speedup compared to existing approaches. Furthermore, we explore the effect of simple thresholding on the generated fields, analyzing the morphological features and power-spectrum characteristics of the resulting two-phase maps. Our results confirm that the proposed framework not only accelerates hyperuniform field synthesis but also provides a versatile platform for systematic study of binary microstructures derived from continuous designs. This work opens new avenues for large-scale simulation and optimized design of advanced hyperuniform materials. |
| title | Ultra-Efficient Reconstruction of Anisotropic Hyperuniform Continuous Random Fields in 2D and 3D via Generalized Spectral Filtering |
| topic | Materials Science |
| url | https://arxiv.org/abs/2509.08675 |