A new web-supervised method for image dataset constructions

Yazhou Yao, Jian Zhang, Fumin Shen, Xiansheng Hua, Jingsong Xu, Zhenmin Tang
2017 Neurocomputing  
The goal of this work is to automatically collect a large number of highly relevant images from Internet for given queries. A novel automatic image dataset construction framework is proposed by employing multiple query expansions. In specific, the given queries are first expanded by searching in Manuscript Click here to view linked References supervised, web supervised and full supervised learning, the experimental results indicated that the effectiveness of our new image dataset construction
more » ... thod is superior to weak supervised and web supervised state-of-the-art methods. In addition, we do a cross-dataset classification to evaluate the performance of our dataset with two publically manual labelled dataset STL-10 and CIFAR-10.
doi:10.1016/j.neucom.2016.07.066 fatcat:stgxtjouyzberhab7htgthk7am