A New Trademark Detection Method via Trademark Confidence Score of MSERs

摘要

This paper proposes a new algorithm to provide high quality potential trademark locations for trademark detection. Inreal-world circumstances, trademark regions often possess some distinctive, invariant and stable properties which can begained effectively and efficiently by Maximally Stable Extremal Regions (MSERs). Based on this observation, wedesign Trademark Confidence Score (TCS) for adaptive MSERs in the images. Then a window refinement algorithm isproposed to retain the high-quality candidate windows generated by Selective Search (SS). Experiments onFlickerLogos-27 and our own dataset demonstrate that our algorithm can significantly reduce the number of candidateproposals produced by SS with little sacrifice of recall for trademarks. Moreover, for trademark detection, our algorithmhas better performance while reducing the computational cost of detection.