Scene40 is a scene dataset which created based on the "VisualGenome" [https://visualgenome.org/] dataset.
Scene40 is a scene dataset that includes indoor and outdoor scenes, and presented in our paper [https://doi.org/10.1007/s13735-022-00246-5]. The dataset contains 8538 images in 40 scene categories, whose scene labels were determined based on location (e.g., airport, park, etc.) or activity (e.g., horse riding, baseball game, etc.) in the image. Some scene categories are object-based and some of them are layout-based. The figure below shows the number of images in each category. There are at least 100 images in each scene category, and the average number of images in each scene category is 213. In this dataset, there is an annotated scene graph and a scene label corresponding to each scene image. The scene graph describes objects and their relationships in the image.
Special thanks to Mrs. Faezeh Soorani (https://github.com/faezehsoorani1997) for her cooperation in preparing this dataset.