Object Detection
Latin
medieval
manuscripts
wjbmattingly's picture
Update README.md
8f5eddd verified
---
license: apache-2.0
datasets:
- CATMuS/medieval-segmentation
language:
- la
base_model:
- Ultralytics/YOLO11
pipeline_tag: object-detection
tags:
- medieval
- manuscripts
---
This a collection of YOLOv11 models trained on the CATMuS Medieval Segmentation dataset which is a collection of various datasets that have [SegmOnto](https://segmonto.github.io) XML markup.
# YOLOv11 Model Evaluation Results
## Overall Metrics
| Model | mAP50-95 | mAP50 | mAP75 |
|-------|-----------|--------|--------|
| YOLOv11-Nano | 0.298 | 0.475 | 0.302 |
| YOLOv11-Small | 0.352 | 0.566 | 0.380 |
| YOLOv11-Medium | 0.368 | 0.576 | 0.392 |
| YOLOv11-Large | 0.389 | 0.599 | 0.405 |
| YOLOv11-XLarge | 0.370 | 0.586 | 0.393 |
## Per-Class mAP50-95 Metrics
| Class | YOLOv11-Nano | YOLOv11-Small | YOLOv11-Medium | YOLOv11-Large | YOLOv11-XLarge |
|--------|-----------|-----------|-----------|-----------|-----------|
| MarginTextZone | 0.179 | 0.234 | 0.243 | 0.202 | 0.251 |
| DefaultLine | 0.727 | 0.738 | 0.758 | 0.763 | 0.759 |
| MainZone | 0.856 | 0.865 | 0.874 | 0.886 | 0.881 |
| RunningTitleZone | 0.390 | 0.465 | 0.461 | 0.497 | 0.398 |
| NumberingZone | 0.112 | 0.158 | 0.136 | 0.136 | 0.121 |
| QuireMarksZone | 0.024 | 0.156 | 0.128 | 0.216 | 0.165 |
| HeadingLine | 0.309 | 0.272 | 0.345 | 0.383 | 0.355 |
| DropCapitalZone | 0.253 | 0.286 | 0.350 | 0.313 | 0.324 |
| StampZone | 0.235 | 0.415 | 0.417 | 0.622 | 0.549 |
| GraphicZone | 0.338 | 0.356 | 0.374 | 0.349 | 0.356 |
| InterlinearLine | 0.032 | 0.075 | 0.128 | 0.118 | 0.122 |
| DigitizationArtefactZone | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |
| DropCapitalLine | 0.120 | 0.200 | 0.203 | 0.184 | 0.158 |
| DamageZone | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |
| MusicLine | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |
| TitlePageZone | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |
| SealZone | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |
| MusicZone | 0.298 | 0.352 | 0.368 | 0.389 | 0.370 |