PatchGraph-MTFormer: A Multitask Patch-Graph Transformer for Hyperspectral Image Analysis
Bibliographic record
Abstract
Hyperspectral image (HSI) classification remains challenging because of the large number of spectral channels, strong spatial-spectral redundancy, and limited labeled samples available in remote sensing problems. In this paper, a graph-structured transformer architecture, named PatchGraph-MTFormer, is presented to address these limitations. The proposed design models the hyperspectral patches as small grid graphs and applies transformer-based attention on graph neighborhoods, enabling simultaneous learning of local spectral signatures and global spatial context. The pipeline integrates grounded preprocessing, including reflective padding, patchbased subdivision, spectral-spatial embedding, and principal component analysis (PCA). PatchGraph-MTFormer is validated on four widely used hyperspectral benchmark datasets, i.e.; Indian Pines, Pavia University, Houston 2013, and WHU-Hi-LongKou and is subsequently extended to multitask plant phenotyping using the HyperLeaf2024 dataset. On the four HSI benchmarks, PatchGraph-MTFormer attains accuracy 99.97% (Indian Pines), 99.47% (Pavia University), 100.00%(Houston 2013), and 99.75% (WHU-Hi-LongKou), respectively, achieving state-of-the-art performance compared with representative classical, CNN-based, graph-based, and transformer-based HSI models. On HyperLeaf2024, the multitask extension achieves a cultivar classification accuracy of 91.5% (macro F1-score ≈ 0.92) and strong regression performance, with a variance-weighted coefficient of determination mean squared error ≈ 0.4, R 2 ≈ 0.593 on standardized targets. Unlike previous hyperspectral methods, PatchGraph-MTFormer unifies local graph topology and transformer attention within patch neighborhoods, enabling context-aware classification while preserving spatial structure. The overall pipeline is to the best of our knowledge, the first to combine multi-scale graph-structured patching, rigorous PCA pooling, and graph-restricted transformer blocks in a multitask setup for both classification and plant phenotyping.
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 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 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".