Labeled Trash Dataset (Pascal VOC, COCO and Yolo)

Termed and Augmented Residue Images File (8 Classes, Pascal VOC, COCO, and Yolo)

• Comprehensive Resource: Explore our meticulously Termed residue File, a cornerstone for researchers and policymakers.
• Innovative Solutions: Harness detailed annotations to develop cutting-edge recycling and waste sorting technologies.
• Environmental Impact: Address pressing challenges by analyzing diverse trash categories, from plastics to organic waste.
• Sustainable Future: Empower data-driven strategies to enhance waste management, promote recycling, and preserve the environment.
• Collaborative Efforts: Join us in utilizing this invaluable resource to drive meaningful change and foster a cleaner, healthier planet.”

This dataset consists of 163 images of trash belonging to 8 classes. The dataset includes:

Original Data: 163 labeled original images
Resized Data: 163 labeled images resized to 1024×1024 pixels
Augmented Data: 4923 images generated by the Burobot library using the Albumentations library
Split Data: Data split into train, test, and validation sets (70% train, 10% test, 20% validation)
Annotation Formats: Label files in Pascal VOC, COCO, and Yolo formats

Classes:

1. Bottle
2. Bottle Cap
3. Can
4. Cardboard
5. Cigarette Box
6. Cigarette Butt
7. Cup
8. Plastic

Additional Information:
Data was augmented using the Burobot library.

This dataset is sourced from Kaggle.