WasteInsight: Dataset for Detection and Volume Estimation of Municipal Solid Waste

The main objective of this dataset is to provide a labeled database to train object detection algorithms capable of identifying trash bags and estimating their volumes based on their dimensions. The dataset contains 447 images, divided into 60% for training, 20% for validation and 20% for testing.

The images were collected in homes, using mobile device cameras, capturing garbage bags in different environments and lighting conditions. Each image in the dataset is accompanied by labels that indicate the coordinates of the bounding boxes around the trash bags.