Feng et al. (2026) MS ² AE-Net: A Multiscale Spectral–Spatial Autoencoder Network for Hyperspectral Unmixing

Identification

• Journal: IEEE Transactions on Geoscience and Remote Sensing
• Year: 2026
• Date: 2026-01-01
• Authors: Ruyi Feng, Hangqi Peng, Lajiao Chen
• DOI: 10.1109/tgrs.2026.3662051

Research Groups

[Not available in the provided text.]

Short Summary

This paper introduces MS2AE-Net, a multiscale spectral-spatial autoencoder network designed for hyperspectral unmixing.

Objective

• To propose and evaluate MS2AE-Net, a novel deep learning architecture, for effective hyperspectral unmixing by leveraging both spectral and spatial information at multiple scales.

Study Configuration

• Spatial Scale: Pixel-level (inherent to hyperspectral unmixing).
• Temporal Scale: Not explicitly applicable; the method is likely applied to static hyperspectral images.

Methodology and Data

• Models used: MS2AE-Net (Multiscale Spectral–Spatial Autoencoder Network).
• Data sources: Hyperspectral imagery (general type, specific sources not provided).

Main Results

[Not available in the provided text.]

Contributions

[Not available in the provided text.]

Funding

[Not available in the provided text.]

Citation

@article{Feng2026MS,
  author = {Feng, Ruyi and Peng, Hangqi and Chen, Lajiao},
  title = {MS <sup>2</sup> AE-Net: A Multiscale Spectral–Spatial Autoencoder Network for Hyperspectral Unmixing},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  year = {2026},
  doi = {10.1109/tgrs.2026.3662051},
  url = {https://doi.org/10.1109/tgrs.2026.3662051}
}