With the explosive growth of commercial Iogos, high quality logo images are needed for training logo detection or recognition systems, especially for famous Iogos or new commercial brands. This paper focuses on automa...With the explosive growth of commercial Iogos, high quality logo images are needed for training logo detection or recognition systems, especially for famous Iogos or new commercial brands. This paper focuses on automatic collecting representative logo images from the internet without any human labeling or seed images. We propose multiple dictionary invariant sparse coding to solve this problem. This work can automatically provide prototypes, representative images, or weak labeled training images for logo detection, logo recognition, trademark infringement detection, brand protection, and ad-targeting. The experiment results show that our method increases the mean average precision for 25 types of Iogos to 80.07% whereas the original search engine results only have 32% representative logo images. The top images collected by our method are accurate and reliable enough for practical applications in the future.展开更多
基金National Key Basic Research and Development(973)Program of China(Nos.2012CB316301 and 2013CB329403)National Natural Science Foundation of China(No.91120011)+1 种基金Tsinghua University Initiative Scientific Research Program(No.20121088071)Basic Research Foundation of Tsinghua National Laboratory for Information Science and Technology(TNList)
文摘With the explosive growth of commercial Iogos, high quality logo images are needed for training logo detection or recognition systems, especially for famous Iogos or new commercial brands. This paper focuses on automatic collecting representative logo images from the internet without any human labeling or seed images. We propose multiple dictionary invariant sparse coding to solve this problem. This work can automatically provide prototypes, representative images, or weak labeled training images for logo detection, logo recognition, trademark infringement detection, brand protection, and ad-targeting. The experiment results show that our method increases the mean average precision for 25 types of Iogos to 80.07% whereas the original search engine results only have 32% representative logo images. The top images collected by our method are accurate and reliable enough for practical applications in the future.
