NUS-wide-128

Introduction

---

NUW-WIDE-128: A Real-World Web Image Database. The dataset is based on NUS-WIDE and includes:

(1) 269,648 images and the associated tags from Flickr, with a total number of 5,018 unique tags.

(2) Six types of low-level features extracted from these images, including 64-D color histogram, 144-D color correlogram, 73-D edge direction histogram, 128-D wavelet texture, 225-D block-wise color moments and 500-D bag of words based on SIFT descriptions.

(3) Ground-truth for 128 concepts that can be used for evaluation.

(4) User information of 247,849 images.

 

Download

ConceptsList.rar (625 downloads )
  
Groundtruth.rar (1058 downloads )
  
Flickr/image data             [baidu]
  
ImageList.rar (951 downloads )
  
imgLabes.rar (1033 downloads )
  
Low_Level_Features         [baidu]
  
NUS_WID_Tags.rar (909 downloads )

 

Citation

---

If you make use of the NUS-WIDE-128 dataset in any form, please cite the following reference.

@article{tang2017tri,
title={Tri-Clustered Tensor Completion for Social-Aware Image Tag Refinement},
author={Tang, Jinhui and Shu, Xiangbo and Qi, Guo-Jun and Li, Zechao and Wang, Meng and Yan, Shuicheng and Jain, Ramesh},
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
year={2017},
doi={10.1109/TPAMI.2016.2608882},
publisher={IEEE}
}

@article{tang2016generalized,
title={Generalized Deep Transfer Networks for Knowledge Propagation in Heterogeneous Domains},
author={Tang, Jinhui and Shu, Xiangbo and and Li, Zechao and Qi, Guo-Jun and Wang, Jingdong},
journal={ACM Transactions on Multimedia Computing, Communications, and Applications},
volume = {12},
number = {4s},
year={2016}
}