Visual Saliency and Text Confidence Map Based Background Suppression for Scene Text
Jiang We · Dianzi xuebao · 2015
The complex background is an important inference factor of scene text automatic localization algorithm. To eliminate the impact of background,this scheme firstly obtains the saliency region in saliency map according to the spectral residual theory. Then it computes text confidence map in saliency region,w hich is w eighted sum of the fraction map of corresponding pairs based on stroke feature and the posterior probability map based on Fourier spectrum statistical feature. Finally,it combines saliency map,text confidence map and HSI color feature w ith the graph model to suppress the background and highlight the foreground text. The experimental results on both the ICDAR2011 scene text localization competition test dataset and laboratory Chinese dataset show that the scheme can effectively suppress the complex background and improve the scene text localization algorithm.