关键词: 文本定位, 二值化, 自然场景图像, 卷积神经网络

Abstract:

Due to the complexity of the natural scene image, it is still difficult to extract the text information from the image, Based on the research and analysis of image text localization algorithm, a localization algorithm based on convolution neural network is proposed. Firstly, the image is preprocessed to obtain the candidate text region，On this basis, combined with convolution neural network to extract the text features for further positioning automatically。Through the experimental verification, the positioning accuracy can reach %86 and has better performance.

Key words: text localization, binarization, natural scene image, convolution neural network