Piece-wise linearity based method for text frame classification in video
作者：
Highlights：
• A new piece-wise linearity based method is proposed for text frame classification.
• New method for text candidate selection has been proposed utilising RGB information.
• The method is capable of handling curved/arbitrarily oriented video text.
• Efforts have been made to overcome the limitation of the state-of-the-art methods.
• Proposed method is computationally less expensive and outperformed existing methods.
摘要
Highlights•A new piece-wise linearity based method is proposed for text frame classification.•New method for text candidate selection has been proposed utilising RGB information.•The method is capable of handling curved/arbitrarily oriented video text.•Efforts have been made to overcome the limitation of the state-of-the-art methods.•Proposed method is computationally less expensive and outperformed existing methods.
论文关键词：Video document processing,Component centroid linearity,Text frame classification,Video text processing,Video text detection
论文评审过程：Received 7 February 2014, Revised 6 August 2014, Accepted 13 September 2014, Available online 22 September 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.09.012
