HSIFormer: An Efficient Vision Transformer Framework for Enhanced Hyperspectral Image Classification Using Local Window Attention
Bibliographic record
Abstract
Convolutional neural networks (CNNs) have recently gained significant attention in image classification due to their exceptional performance in computer vision. Building on this success, researchers are now investigating the potential of transformers in Earth observation applications. However, transformers face a significant challenge: they require substantially more training data than CNN classifiers. This makes their application in remote sensing, particularly with Hyperspectral Image (HSI) data, difficult due to the limited availability of labeled data. In this paper, we will repurpose the PolSARFormer model for hyperspectral image classification. Originally designed for polarized SAR image classification, the model's initial parameters have been fine-tuned to better meet the requirements of hyperspectral data. The PolSARFormer model employs a vision transformer (ViT)-based framework that utilizes 3D and 2D CNNs as feature extractors and incorporates local window attention (LWA) for effective HSI data classification. Extensive experimental results show that the model, HSIFormer, achieves better classification accuracy than the state-of-the-art Swin Transformer and ViT algorithms. HSIFormer outperformed the Swin Transformer and ViT by 2.31% and 3.24% in overall accuracy (OA) on the Pavia University benchmark dataset. Additionally, results on the Salinas dataset demonstrated that HSIFormer surpasses several other algorithms, including HybridSN (96.89%), Tri-CNN (97.05%), Vision Transformer (94.68%), 3D-CNN (96.77%), and Swin Transformer (95.62%), with a kappa index (KI) of 98.23%. The code will be made publicly available at https://https://github.com/mqalkhatib/HSIFormer
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".