metadata
license: mit
task_categories:
- image-segmentation
language:
- en
pretty_name: Jigsaw Puzzles
size_categories:
- n<1K
Images of various jigsaw puzzle pieces gathered from a website called Jigsaw Explorer.
Segmentations were generated by SAM 3 with some image preprocessing. Details about how I created this dataset can be found in this github repository
Follows the schema used by YOLO segmentation models, images are resized to 1920x1080