A Hangeul Recognition Method Using Directional Edges in Open Captions

Abstract

This paper proposes an efficient method to recognize Hangeul in video open captions. The open captions in news video can play an important role in the video indexing. The strokes of Korean character have a very strong horizontal and vertical directionality and some strokes appear repeatedly in each character. Based on this characteristics, in this paper, we propose an efficient algorithm to extract the character region in open caption and recognize the characters based on these characteristics of Korean character. The simulation results demonstrate the efficiency of our algorithm in terms of computation time and recognition accuracy.