YOLOV8-CMS: a high-accuracy deep learning model for automated citrus leaf disease classification and grading
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: Citrus leaf diseases significantly affect production efficiency and fruit quality in the citrus industry. To effectively identify and classify citrus leaf diseases, this study proposed a classification approach leveraging deep learning techniques (YOLOV8 equipped with CSPPC, MultiDimen, SpatialConv, YOLOV8-CMS). Additionally, a segmentation method was utilized to extract leaf and lesion areas for disease severity grading based on their pixel ratio. RESULTS: By collecting and preprocessing a citrus leaf image dataset, the YOLOV8-CMS model was trained for disease classification. The model integrated MultiDimen attention, SpatialConv, and the CSPPC module to enhance performance. Furthermore, a segmentation approach was applied to precisely segment both leaf and lesion areas, enabling a quantitative assessment of disease severity. To verify the effectiveness of the proposed approach, multiple YOLO-based architectures, including different YOLOV8 series models, YOLOV5, and YOLOV3, were compared and analyzed. Results demonstrated that the proposed method achieved outstanding performance in citrus leaf disease classification, with an mAP50 of 98.2% in distinguishing healthy and diseased leaves and an accuracy of 97.9% in multi-class disease classification tasks. CONCLUSIONS: The proposed YOLOV8-CMS model outperformed traditional methods in citrus leaf disease classification, while the segmentation-based approach enabled an accurate and quantitative assessment of disease severity. These findings highlighted the potential of deep learning in precision agriculture, contributing to more effective disease management in citrus production.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it