DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder
Hyperspectral image (HSI) classification is essential for ecological monitoring, but faces significant challenges due to high dimensionality, complex spectral–spatial relationships, and limited labeled data. This study introduces DiffusionAAE, a novel framework that uniquely combines Adversarial Aut...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | Ecological Informatics |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S157495412500127X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850201346403205120 |
|---|---|
| author | Zeyu Cao Jinhui Li Xiangrui Xu |
| author_facet | Zeyu Cao Jinhui Li Xiangrui Xu |
| author_sort | Zeyu Cao |
| collection | DOAJ |
| description | Hyperspectral image (HSI) classification is essential for ecological monitoring, but faces significant challenges due to high dimensionality, complex spectral–spatial relationships, and limited labeled data. This study introduces DiffusionAAE, a novel framework that uniquely combines Adversarial Autoencoders (AAE) with conditional diffusion models to address these challenges. Unlike existing approaches, DiffusionAAE incorporates spectral similarity constraints and class label guidance into the diffusion process, ensuring the generation of physically realistic synthetic samples. Our framework’s key innovation lies in its two-stage architecture: first, an AAE extracts robust latent features capturing intricate spectral–spatial relationships; second, a conditional diffusion model refines these features through progressive denoising, enabling class-specific feature generation while maintaining physical constraints inherent to hyperspectral data. Comprehensive experiments on three benchmark datasets demonstrate DiffusionAAE’s superior performance: compared to state-of-the-art methods, our approach achieves significant improvements with an overall accuracy (OA) of 96.77% on Indian Pines (3.21% higher than CNN-based methods), 99.56% on University of Pavia (1.24% improvement over Transformer-based approaches), and 99.62% on Salinas (0.98% better than the best competing method). Notably, DiffusionAAE shows remarkable performance on minority classes, with an average 7.35% accuracy improvement across underrepresented classes in the Indian Pines dataset. The framework demonstrates particular strength in scenarios with limited training data, maintaining 95.3% accuracy even when using only 5% of available labeled samples. These results establish DiffusionAAE as a significant advancement for ecological informatics applications, especially for biodiversity monitoring and land cover classification where labeled data scarcity and class imbalance are prevalent challenges. |
| format | Article |
| id | doaj-art-ad41b692c5984e31b024c955e44b6c99 |
| institution | OA Journals |
| issn | 1574-9541 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Ecological Informatics |
| spelling | doaj-art-ad41b692c5984e31b024c955e44b6c992025-08-20T02:12:02ZengElsevierEcological Informatics1574-95412025-07-018710311810.1016/j.ecoinf.2025.103118DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial AutoencoderZeyu Cao0Jinhui Li1Xiangrui Xu2School of Spatial Planning and Design, Hangzhou City University, Hangzhou, 310015, China; Research Institute for Spatial Planning and Design, Hangzhou City University, Hangzhou, 310015, ChinaCollege of Electrical Engineering, Zhejiang University, Hangzhou, 310027, ChinaSchool of Spatial Planning and Design, Hangzhou City University, Hangzhou, 310015, China; Research Institute for Spatial Planning and Design, Hangzhou City University, Hangzhou, 310015, China; Correspondence to: 51 Huzhou Street, Hangzhou, Zhejiang, 310015, China.Hyperspectral image (HSI) classification is essential for ecological monitoring, but faces significant challenges due to high dimensionality, complex spectral–spatial relationships, and limited labeled data. This study introduces DiffusionAAE, a novel framework that uniquely combines Adversarial Autoencoders (AAE) with conditional diffusion models to address these challenges. Unlike existing approaches, DiffusionAAE incorporates spectral similarity constraints and class label guidance into the diffusion process, ensuring the generation of physically realistic synthetic samples. Our framework’s key innovation lies in its two-stage architecture: first, an AAE extracts robust latent features capturing intricate spectral–spatial relationships; second, a conditional diffusion model refines these features through progressive denoising, enabling class-specific feature generation while maintaining physical constraints inherent to hyperspectral data. Comprehensive experiments on three benchmark datasets demonstrate DiffusionAAE’s superior performance: compared to state-of-the-art methods, our approach achieves significant improvements with an overall accuracy (OA) of 96.77% on Indian Pines (3.21% higher than CNN-based methods), 99.56% on University of Pavia (1.24% improvement over Transformer-based approaches), and 99.62% on Salinas (0.98% better than the best competing method). Notably, DiffusionAAE shows remarkable performance on minority classes, with an average 7.35% accuracy improvement across underrepresented classes in the Indian Pines dataset. The framework demonstrates particular strength in scenarios with limited training data, maintaining 95.3% accuracy even when using only 5% of available labeled samples. These results establish DiffusionAAE as a significant advancement for ecological informatics applications, especially for biodiversity monitoring and land cover classification where labeled data scarcity and class imbalance are prevalent challenges.http://www.sciencedirect.com/science/article/pii/S157495412500127XLand cover classificationGenerative modelFeature extractionHyperspectral image classificationConditional diffusion modelAdversarial autoencoder |
| spellingShingle | Zeyu Cao Jinhui Li Xiangrui Xu DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder Ecological Informatics Land cover classification Generative model Feature extraction Hyperspectral image classification Conditional diffusion model Adversarial autoencoder |
| title | DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder |
| title_full | DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder |
| title_fullStr | DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder |
| title_full_unstemmed | DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder |
| title_short | DiffusionAAE: Enhancing hyperspectral image classification with conditional diffusion model and Adversarial Autoencoder |
| title_sort | diffusionaae enhancing hyperspectral image classification with conditional diffusion model and adversarial autoencoder |
| topic | Land cover classification Generative model Feature extraction Hyperspectral image classification Conditional diffusion model Adversarial autoencoder |
| url | http://www.sciencedirect.com/science/article/pii/S157495412500127X |
| work_keys_str_mv | AT zeyucao diffusionaaeenhancinghyperspectralimageclassificationwithconditionaldiffusionmodelandadversarialautoencoder AT jinhuili diffusionaaeenhancinghyperspectralimageclassificationwithconditionaldiffusionmodelandadversarialautoencoder AT xiangruixu diffusionaaeenhancinghyperspectralimageclassificationwithconditionaldiffusionmodelandadversarialautoencoder |