datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
pcb-detection-augmented-obb
PCB Detection Augmented OBB Dataset
Augmented PCB oriented-bounding-box data in YOLO format from
SanderGi/PCB-Detection.
The repository includes train, validation, and test images; OBB labels; dataset
configuration; and training outputs used by the published PCB OBB model.
Download
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="SanderGi/pcb-detection-augmented-obb",
repo_type="dataset",
local_dir="data/augmented_obb",
)
See… See the full description on the dataset page: https://huggingface.co/datasets/SanderGi/pcb-detection-augmented-obb.YOLO-OBB-Zebra-Crossings-Datasetcube-detection-obb
Colored Cubes OBB Detection Dataset
A small object-detection dataset for oriented bounding box (OBB) detection of
four colored cubes (green, yellow, blue, red). Intended for training and
benchmarking YOLO-OBB style models in robotic-manipulation and pick-and-place
contexts.
Dataset Summary
Task: Oriented bounding box detection (4-point polygon per object)
Classes: 4 — green_cube, yellow_cube, blue_cube, red_cube
Images: 215 total · 1280×720 JPEG
Format: Ultralytics… See the full description on the dataset page: https://huggingface.co/datasets/movensys/cube-detection-obb.cube-detection-monoply-background-obb
Cube Detection on Monopoly Board Background (OBB)
A small oriented-bounding-box (OBB) detection dataset of colored cubes placed on a Movensys "Monopoly" board background. Intended for fine-tuning YOLO-style OBB detectors used in pick-and-place / robotic manipulation pipelines.
Classes
ID
Name
0
green_cube
1
yellow_cube
2
blue_cube
3
red_cube
Splits
Split
Images
Labels
train
104
104
val
29
29
test
16
16
total
149
149… See the full description on the dataset page: https://huggingface.co/datasets/movensys/cube-detection-monoply-background-obb.dice-detection-obb
Dice Detection (OBB) with Pip Classification
A small oriented-bounding-box (OBB) detection dataset of a wooden die captured from a top-down camera. Each annotation gives the die's rotated bounding box along with the visible pip count (1–6) as the class. Intended for fine-tuning YOLO-style OBB detectors used in pick-and-place / robotic manipulation pipelines.
Classes
ID
Name
0
Five
1
Four
2
One
3
Six
4
Three
5
Two
(Class IDs are not in numerical… See the full description on the dataset page: https://huggingface.co/datasets/movensys/dice-detection-obb.rwy_obb-300-65-65_v2
Runway Object-Oriented Bounding Box (OBB) Segmentation Dataset v2
Dataset Description
This dataset contains aerial/satellite imagery for runway segmentation using oriented bounding boxes (OBB). The dataset is specifically designed for semantic segmentation tasks focusing on runway detection and delineation in aerial imagery.
Dataset Statistics
Split
Images
Labels
Train
303
303
Val
65
65
Test
65
65
Classes
{
"0": "_background_"… See the full description on the dataset page: https://huggingface.co/datasets/Spatiallysaying/rwy_obb-300-65-65_v2.rubiks-cube-obb
Rubik's Cube Oriented Bounding Box Detection
A dataset of 148 images (143 train / 5 val) of Rubik's cubes annotated with
oriented bounding boxes (OBB) for color and face detection.
Classes
ID
Name
0
Blue
1
Green
2
Orange
3
Red
4
White
5
Yellow
6
cube_face
7
side_face
Note: class 7 (side_face) is effectively unused (only 1 instance in training).
Annotation Format
Each image has a corresponding list of objects. Each… See the full description on the dataset page: https://huggingface.co/datasets/Alabi-Ayobami/rubiks-cube-obb.rwy_obb_mask2former-300-65-65
Runway Object-Oriented Bounding Box (OBB) Mask2Former Dataset v2
Dataset Description
This dataset contains aerial/satellite imagery for runway segmentation using Mask2Former architecture. The dataset supports both semantic and instance segmentation tasks, specifically designed for runway detection and delineation in aerial imagery.
Key Features
Architecture: Optimized for Mask2Former universal segmentation
Task Support: Semantic, instance, and panoptic… See the full description on the dataset page: https://huggingface.co/datasets/Spatiallysaying/rwy_obb_mask2former-300-65-65.robotic-garbage-obb
Robotic Garbage OBB Dataset (robotic-garbage-obb)
This dataset is designed for robotic parallel gripper grasping and top-down garbage classification using Oriented Bounding Boxes (OBB).
Images are captured in a real robotic bird's-eye view workspace (pure black background tabletop) using Intel RealSense cameras, labeled with oriented bounding boxes to provide both classification and exact planar yaw ($\theta$) orientation angles for robotic parallel grippers.
📸… See the full description on the dataset page: https://huggingface.co/datasets/0326BlueW/robotic-garbage-obb.license-detection-paligemma-obbobb-303-65-65
Spatiallysaying/obb-303-65-65
This dataset provides deterministic splits derived from Spatiallysaying/obb.
Source & Method
Source splits: train=346, validation=87 (total 433)
New splits: train=303, validation=65, test=65
Policy: test=15% of total; validation=15% of total carved from the remaining 85%
Seed: 1337
Procedure: using 🧡 Datasets train_test_split twice for deterministic partitioning.
Notes
Task: Semantic segmentation (image + mask).
Background… See the full description on the dataset page: https://huggingface.co/datasets/Spatiallysaying/obb-303-65-65.cotton-fisheye-obb