Lightweight Vision Transformer with Cross-Scale Attention Fusion for Field-Deployable Maize Disease Diagnosis
Bibliographic record
Abstract
Early and accurate detection of maize leaf diseases is crucial for preventing yield loss and reducing agrochemical use. Existing vision-based models struggle in realworld agricultural conditions due to data imbalance, lighting variability, and high computational demands. This paper introduces MaizeFormerX, a lightweight Vision Transformer that utilizes multi-scale patch embedding and a Cross-Scale Attention Fusion (CSAF) module. This design enables effective detection of fine-grained lesion textures and broader disease patterns while maintaining a low parameter count. MaizeFormerX was evaluated on two public datasets (Dataverse and Tanzania) using specific preprocessing, targeted augmentation, and stratified $\mathbf{1 0}$-fold cross-validation. The model achieved accuracies of $97.8 \%, 97.5 \%$, and 96.9% on the Dataverse, Tanzania, and Plagues Maiz datasets, respectively, outperforming the Swin Transformer V2 by $2-3 \%$ with significantly fewer parameters and lower computational costs. Grad-CAM visualizations provide pixel-level interpretability, and this has been implemented in a cost-effective web application for real-time field use. This research offers an accurate, efficient, and explainable framework for diagnosing maize diseases, bridging academic findings with practical agricultural applications, and has the potential to enhance crop management in resource-constrained environments.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".