LayerWatch
A Benchmark for Layer-wise Multi-Class Visual Defect Detection in Fused Filament Fabrication (FFF).
Dataset Description
This dataset supports visual defect detection and classification in Fused Filament Fabrication (FFF) 3D printing.
Original Sources
Harvard Dataverse: https://doi.org/10.7910/DVN/AWWBAZ
Project repository: https://github.com/SEilermann/LayerWatch
Dataset Statistics
- Images: 28,018
- Label files: 28,013
- Empty annotation files: 245
- Missing annotation files: 5
- Classes: 6
- Image resolution: 640x640
Classes
- CloggedNozzle
- Delamination
- OK
- LayerShift
- OffPlatform
- Warping
Dataset Structure
The original directory structure is preserved.
data/ train/ train/ images/ labels/ valid/ images/ labels/ benchy/ valid/ images/ labels/ clamp-part_o/ valid/ images/ labels/ clamp-part_s/ valid/ images/ labels/ pyramide/ valid/ images/ labels/
Each dataset directory contains its corresponding data.yaml configuration.
Annotation Format
YOLO object detection format:
class_id x_center y_center width height
Coordinates are normalized.
Audit Notes
The original dataset contains 28,018 images and 28,013 label files.
Five images were identified without corresponding label files during preprocessing.
Empty annotation files are preserved.
License and Attribution
Original dataset license: CC BY-NC 4.0.
Please refer to the original Harvard Dataverse record for licensing terms and attribution.
Original dataset content and annotations are preserved without intentional modification.
- Downloads last month
- 64