基于内容风格迁移的残缺稀疏文字图像盲修复方法  

Blind Restoration Method for Incomplete and Sparse Text Images Based on Content Style Transfer

作　　者：刘栋斌 王慧琴[1] 王可[1] 王展 甄刚 Liu Dongbin;Wang Huiqin;Wang Ke;Wang Zhan;Zhen Gang(School of Information and Control Engineering,Xi’an University of Architecture and Technology,Xi’an 710055,Shaanxi,China;Shaanxi Provincial Institute of Cultural Relics Protection,Xi’an 710075,Shaanxi,China)

机构地区：[1]西安建筑科技大学信息与控制工程学院,陕西西安710055 [2]陕西省文物保护研究院,陕西西安710075

出　　处：《激光与光电子学进展》2022年第24期98-109,共12页Laser & Optoelectronics Progress

基　　金：陕西省自然科学基础研究计划项目(2021JM-377);陕西省科技厅科技合作项目(2020KW-012);陕西省教育厅智库项目(18JT006);西安市科技局高校人才服务企业项目(GXYD10.1)。

摘　　要：现有残缺文字图像的修复需要确定Mask区域后对其填补,如果文字残缺部分剩余信息量过于稀疏,将无法确定Mask区域。针对该问题,提出了一种基于内容风格迁移的残缺稀疏文字图像盲修复方法。利用循环生成对抗网络构建修复前后文字图像间的全局关联像素信息,将残缺文字的图像内容风格特征迁移为完整文字图像从而进行修复;并在网络中加入自注意力机制对稀疏像素进行全局约束,解决迁移过程中相隔较远文字稀疏像素之间依赖关系较弱的问题;同时在自注意力机制中使用最大池化,提高迁移修复后的文字图像纹理特征;使用最小二乘损失替换原网络模型中的sigmoid交叉熵损失函数,提高迁移精度。实验结果表明,所提方法不借助Mask指导,能够盲修复稀疏性残缺文字图像中的随机未知缺失区域。The existing incomplete text image restoration must identify and fill the Mask region.The Mask region cannot be determined if the residual information of the incomplete text part is too sparse.A blind restoration method of incomplete and sparse text images based on content style transfer is proposed to address this issue.Using a cyclic generative adversarial network to construct the global related pixel information between the text images before and after restoration,the image content style features of the incomplete text were transferred to complete text images for restoration.The selfattention mechanism is added to the network to globally restrict the sparse pixels,thereby resolving the issue of weak dependence between the text sparse pixels far away in the migration process.simultaneously,the maximum pooling is used in the selfattention mechanism to enhance the texture features of the text images after migration and restoration.To improve migration accuracy,the least square loss function is used to replace the sigmoid cross-entropy loss function in the original network model.The proposed algorithm can repair random unknown missing regions in sparse incomplete text images without the assistance of Mask,according to experimental results.

关 键 词：成像系统 盲修复 循环生成对抗网络 自注意力模块 最小二乘损失 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]