Text/non-text image classification in the wild with convolutional neural networks

作者：

Highlights：

• We study a new and important problem: text/non-text image classification in the wild.

• A new scheme based on block-level classification is proposed to tackle this problem.

• We propose MSP-Net, a novel CNN variant, to efficiently classify text/non-text images.

• As a by-product, MSP-Net outputs coarse locations and scales of texts.

摘要

Highlights•We study a new and important problem: text/non-text image classification in the wild.•A new scheme based on block-level classification is proposed to tackle this problem.•We propose MSP-Net, a novel CNN variant, to efficiently classify text/non-text images.•As a by-product, MSP-Net outputs coarse locations and scales of texts.

论文关键词：Natural images,Text/non-text image classification,Convolutional neural network,Multi-scale spatial partition

论文评审过程：Received 13 March 2016, Revised 5 December 2016, Accepted 8 December 2016, Available online 11 December 2016, Version of Record 12 March 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.12.005