Authors: Hyung Il Koo and Yong Gyun Kim

Description: This text detection system was submitted by Hyung Il Koo from Ajou University, Suwon, Korea. The method first extracts connected components (CCs) with the maximally stable extremal region (MSER) algorithm. These extracted CCs are grouped into text-lines based on their geometric and photometric similarities. In order to find text-lines in arbitrary orientations, the method also adopts the projection profile analysis. Finally, non-text components are filtered out by using the method in [1].

[1] Hyung Il Koo and Duck Hoon Kim, "Scene Text Detection via Connected Component Clustering and Nontext Filtering," IEEE Transactions on Image Processing, vol.22, no.6, pp.2296,2305, June 2013