A new method for multi-oriented graphics-scene-3D text classification in video

作者：

Highlights：

• We propose a novel method for classifying graphics-scene and 2D–3D texts in video.

• An iterative procedure to identify text candidates is presented.

• Stroke width and medial axis are explored for classifying graphics and scene texts.

• Gradient directions and medial axis are combined for classifying 2D and 3D texts.

摘要

•We propose a novel method for classifying graphics-scene and 2D–3D texts in video.•An iterative procedure to identify text candidates is presented.•Stroke width and medial axis are explored for classifying graphics and scene texts.•Gradient directions and medial axis are combined for classifying 2D and 3D texts.

论文关键词：Text detection,Text recognition,Stroke with distance,Ring radius transform,Graphics and scene text,2D and 3D texts

论文评审过程：Received 22 August 2014, Revised 3 April 2015, Accepted 1 July 2015, Available online 10 July 2015, Version of Record 28 September 2015.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.07.002