A DATASET AT A GLANCE...

Tons of wastes, be it of biodegradable or non-biodegradable nature, are dumped on land and water bodies everyday by industries, educational institutions, organizations, food processing as well as delivering units, hospitals and household or domestic activities. Without proper disposal of these wastes, the environment is greatly polluted and turns harmful that it creates room for new diseases, affecting the existence of life on Earth. Many thanks to the finding of waste treatment and recycling methods, which could transform the waste into a useful product.  Hence, wastes are not going to be just wasted!!! However, the success of waste reuse or recycling relies on the prior classification and detection of waste as plastic, paper, metal and so on. The researchers intending to propose approaches for performing automatic waste classification, prior to recycling, reuse or recovery, can use the TrashBox dataset from the GitHub repository (https://github.com/nikhilvenkatkumsetty/TrashBox). With this dataset, about 17785 waste object images can be classified as that belonging to one of the seven classes viz. medical waste, e- waste, plastic, paper, metal, glass and cardboard. The available classes too have subclasses to differentiate between various trash kinds. The inclusion of medical waste and e-waste is an added feature of this dataset, while comparing with the rest of the available datasets for waste classification.  There is a citation request to use this dataset and it is: ‘N. V. Kumsetty, A. Bhat Nekkare, S. K. S. and A. Kumar M., "TrashBox: Trash Detection and Classification using Quantum Transfer Learning," 2022 31st Conference of Open Innovations Association (FRUCT), 2022, pp. 125-130, Doi: 10.23919/FRUCT54823.2022.9770922’. 

Image Courtesy: www.freepik.com