CAJ | 학술논문

提出了一种基于CDF9-7小波和自适应Otsu算法的视频图像字幕分割算法。首先从视频中截取视频图像，并对其进行灰度化等预处理，其次对预处理后的图像进行CDF9-7小波变换，获取其水平和垂直的高频分量HH，然后使用自适应Otsu算法分割出该高频分量图像中的字幕区域。实验表明，该算法分割效果良好，具有一定的鲁棒性和自适应性。以这些区域为基础，可以进一步完成诸如视频字幕识别等图像分析、理解和识别操作。
제출료일충기우CDF9-7소파화자괄응Otsu산법적시빈도상자막분할산법。수선종시빈중절취시빈도상，병대기진행회도화등예처리，기차대예처리후적도상진행CDF9-7소파변환，획취기수평화수직적고빈분량HH，연후사용자괄응Otsu산법분할출해고빈분량도상중적자막구역。실험표명，해산법분할효과량호，구유일정적로봉성화자괄응성。이저사구역위기출，가이진일보완성제여시빈자막식별등도상분석、리해화식별조작。
A video image subtitle segmentation algorithm based on CDF9-7 wavelet and adaptive Otsu algorithm is proposed in this paper. First,a video image is extracted from the video stream,its pretreatment such as graying is made,the pretreated image′s horizontal and vertical high-frequency component HH is obtained by CDF9-7 wavelet transformation,and then the image subtitle regions in the high frequency component image are segmented by using adaptive Otsu algorithm. Experiments show that the method works well and has a certain robustness and adaptability. Based on these regions,further image analysis,understand-ing and recognition tasks such as video subtitle recognition can be achieved.