Chinese text extraction for video scene by multiframe corner matching and heuristic rules

A Chinese text extraction method for video scene is proposed in this paper. Multi-frame corner matching and heuristic rules are combined to detect the text region candidates roughly, which not only solves the issue of Harris corner filtration for complex background video scene, but also improves the detection accuracy using multi-frame fusion. Local texture description is then used for similarity evaluation judged by SVM. Experimental results for 4 different types of 312-frame video images show the effectiveness of the proposed method compared with 5 existing text extraction methods.