Fig. 14From: Detection and recognition of cursive text from video framesImperfect localization of text regionsBack to article page