[AISp2-1] Extraction of Text Areas in Scene Images by Focusing on Local Contour Density and Corner Pixel Ratio
Neural network, Text area extraction, Scene image, Local contour density, Corner pixel ratio
Extraction of character regions is necessary in various fields such as automatic driving. In this paper, we propose a method for extracting character regions from scene images. The method identifies characters by a neural network using local contour density and corner pixel ratio. The experimental results show that the proposed method obtains sufficient accuracy for character region extraction.