WebLogo-2M: Scalable Logo Detection by Deep Learning From the Web

Hang Su, Shaogang Gong, Xiatian Zhu; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 270-279

Abstract


Existing logo detection methods usually consider a small number of logo classes and limited images per class with a strong assumption of requiring tedious object bounding box annotations, thus unscalable to real-world applications. This work tackles these challenges by exploring the webly data learning principle without exhaustive manual labelling. Specifically, we propose a novel incremental learning approach, called Scalable Logo Self-Training (SLST), capable of automatically self-discovering informative training images from noisy web data for progressive model update. Moreover, we introduce a very large (1,867,177 images of 194 classes) logo dataset "WebLogo-2M" by an automatic web data collection and processing method. Extensive evaluations demonstrate the superiority of the SLST method over state-of-the-art strongly and weakly supervised detection models and webly data learning alternatives.

Related Material


[pdf]
[bibtex]
@InProceedings{Su_2017_ICCV,
author = {Su, Hang and Gong, Shaogang and Zhu, Xiatian},
title = {WebLogo-2M: Scalable Logo Detection by Deep Learning From the Web},
booktitle = {Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2017}
}