NUUR-Texture500: A Diverse Dataset of High Resolution Homogeneous Textures

"NUUR-Texture500" is a dataset of diverse texture images created to facilitate research on texture analysis and synthesis. The textures in this dataset are spatially homogeneous, ranging from regular to stochastic, typically containing repeated elements with random variations in position, shape, orientation and color. For a detailed description of the dataset construction and contents, readers should refer to the following paper: Jue Lin, Gaurav Sharma, Thrasyvoulos N. Pappas, "Towards Universal Texture Synthesis by Combining Texton Broadcasting with Noise Injection in StyleGAN-2", Journal of e-Prime - Advances in Electrical Engineering, Electronics and Energy (3) (2023), https://doi.org/10.1016/j.prime.2022.100092 Permission to copy and use this dataset for noncommercial use is hereby granted provided this notice is retained in all copies and the dataset distribution and the paper mentioned below are clearly cited. Contacts: Jue Lin: [email protected] Gaurav Sharma: [email protected] Thrasyvoulos N. Pappas: [email protected] Disclaimer: The dataset is provided "as is" with ABSOLUTELY NO WARRANTY expressed or implied. Use at your own risk. Acknowledgment: The NUUR-Texture500 texture images are curated from a number of publicly accessible sources. We acknowledge and thank the original sites for their contributions: Flickr www.flickr.com NeedPix www.needpix.com Pexels www.pexels.com PickUpImage www.pickupimage.com Pixabay www.pixabay.com PublicDomainPictures www.publicdomainpictures.net RawPixel www.rawpixel.com Unsplash www.unsplash.com WikimediaCommons commons.wikimedia.org