/TGRS-daan

demo of DAAN

demo of DAAN

DAAN: A Deep Autoencoder-based Augmented Network for Multilinear Hyperspectral Unmixing

Note: DAAN codes are intended for academic communication only and may not be used commercially.

This demo of DAAN was made by Zhiqing Zhu and Yuanchao Su in July 2023.

Contact: Yuanchao Su (suych3@xust.edu.cn) and Lianru Gao (gaolr@aircas.ac.cn).

Architectures and Systems

  • DAAN can run under X86 architectures.
  • The demo was tested under Windows 11.
    #---------------------------------------------------------------------------------------------
    ├── Readme.md
    ├── Environment: Python 3.8.10 and PyTorch 2.0.1
    ├── Demo
    │ └── DAAN.py
    ├── Test data: real and synthetic data sets.
    │ ├── data_real(Real data)
    │ │ ├──Y.npy(Moffett Field data)
    │ │ └──E.npy(Endmembers)
    │ └── data_synthetic(Synthetic data)
    │ ├──Y.npy(Synthetic data)
    │ └──E.npy(Endmembers)
    └── Results files

Specification

  • 'demo_synthetic_data.py' uses synthetic hyperspectral data.
  • 'demo_real_data.py' uses Moffett Field data.