Scene Dataset

Published: 25 October 2026| Version 1 | DOI: 10.17632/r777ck2sfx.1
Contributor:
Altaf Khan

Description

12-Scene image Dataset The dataset is constructed by following the image scene geometries. It contains 12 different categories, namely sky-background-ground (skyBkgGnd), background ground (bkgGnd) sky-ground (skyGnd), table-personbackground (tabPersonBkg), person-background (personBkg), box, ground-diagonal background (groundDiagBkgRL), ground, diagonal background (diagBkgRL), 1 side wall (sidewallRL), corner, and no depth (noDepth). Each category represents a particular 3D rough structure of an image scene. Dataset contains 12000 images in total where each category contains equal number of images. https://www.semanticscholar.org/paper/Texture-Gradient-and-Deep-Features-Fusion-Based-Khan-Chefranov/46bf45169a1a13e570d6df2a9963d3fd3f6d2082 cite our paper: @article{Khan2020TextureGA, title={Texture Gradient and Deep Features Fusion-Based Image Scene Geometry Recognition System Using Extreme Learning Machine}, author={Altaf H. Khan and Alexander G. Chefranov and Hasan Demirel}, journal={2020 3rd International Conference on Intelligent Robotic and Control Engineering (IRCE)}, year={2020}, pages={37-41}, url={https://api.semanticscholar.org/CorpusID:221848464} } The 3rd IEEE International Conference on Intelligent Robotics and Control Engineering (IRCE 2020), 10-12 August, 2020, University of Oxford, UK, DOI: 10.1109/IRCE50905.2020.9199253 Terms and conditions to use the dataset *Dataset is available for only education and research purposes (non-commercial). Note: Images are taken from Places365 dataset Zhou, B., Lapedriza, A., Khosla, A., Oliva, A., & Torralba, A. (2018). Places: A 10 Million Image Database for Scene Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40(6), 1452-1464. doi:10.1109/TPAMI.2017.2723009. Images are annoted by following the scene geometries. For more information, please contact: altaf.khan@umw.edu.pk,

Files

Categories

Image Database, Image Analysis, Scene Understanding

Licence