can you tell me which datasets are
standard ones used in most papers
The most commonly used standard datasets in trash and waste classification research
are:
TrashNet Dataset:
Contains approximately 2,527 images across six classes: glass, paper, cardboard,
plastic, metal, and trash. It is widely referenced and available for download via
GitHub. Many papers utilize this dataset for benchmarking and model training [1][2].
Garbage Classification Dataset (Papers With Code):
Comprises around 15,150 images categorized into 12 classes such as paper,
cardboard, biological waste, metal, plastic, and various types of glass, along with
other household waste categories. It is used for more detailed classification tasks [3].
Garbage Image Dataset (IEEE DataPort):
Consists of images of garbage items collected from localities, categorized into five
classes, collected via smartphones. It is suitable for real-world garbage classification
research[4].
TrashBox Dataset:
Contains about 17,785 images of waste objects from web scraping, covering
categories like medical waste, e-waste, plastic, paper, metal, glass, and cardboard.
It supports multi-class waste classification research [1].
Kaggle Garbage Dataset (V2):
Features approximately 19,762 images across 10 classes, including metal, glass,
biological waste, paper, battery, trash, cardboard, shoes, clothes, and plastic. It is
designed for recycling and waste management applications [1].
DeepSeaWaste Dataset:
Comprises around 3,000 underwater images categorized into four waste types,
useful for marine waste detection[1].
MJU-Waste Dataset:
Contains 2,475 RGB-D image pairs of waste items collected on a university campus,
annotated in PASCAL VOC and COCO formats, focusing on waste in a controlled
environment[1].
Other Notable Datasets:
o Spotgarbage-GINI: 2,561 images with bounding box annotations for trash
detection[1].
o Wadaba Plastic Waste Database: 4,000 images with detailed plastic waste
descriptions[1].
In summary, TrashNet and Garbage Classification Dataset are the most widely
adopted standard datasets in research, providing a solid benchmark for developing and
evaluating waste classification models.
1. https://github.com/AgaMiko/waste-datasets-review
2. https://dl.acm.org/doi/fullHtml/10.1145/3574318.3574345
3. https://paperswithcode.com/dataset/garbage-classification-dataset
4. https://ieee-dataport.org/documents/garbage-image-dataset
