Blind text images deblurring based on a generative adversarial network. Issue 14 (31st October 2019)

• Record Type: Journal Article
• Authors: Qi, Qing; Guo, Jichang
• Is Part Of: IET image processing. Volume 13:Issue 14(2019)
• Page Start: 2850
• Page End: 2858
• Publication Date: 2019-10-31
• Subjects: learning (artificial intelligence) -- text analysis -- image restoration; blind text images -- generative adversarial network -- text images deblurring -- hand‐crafted priors -- specific kernel -- text deblurring problem -- semantic generation task -- structural loss function -- detailed loss function -- multiscale generator -- generated text images -- realistic latent images -- real‐world blurry images; Image processing -- Periodicals; 621.36705
• DOI: 10.1049/iet-ipr.2018.6697
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract : Recently, text images deblurring has achieved advanced development. Unlike previous methods based on hand‐crafted priors or assume specific kernel, the authors recognise the text deblurring problem as a semantic generation task, which can be achieved by a generative adversarial network. The structure is an essential property of text images; thus, they propose a structural loss function and a detailed loss function to regularise the recovery of text images. Furthermore, they learn from the coarse‐to‐fine strategy and present a multi‐scale generator, which is utilised for sharpening the generated text images. The model has a robust capability of generating realistic latent images with photo‐quality effect. Extensive experiments on the synthetic and real‐world blurry images have shown that the proposed network is comparable to the state‐of‐the‐art methods.