福建电脑2024，Vol.40Issue(4)：45-48,4.DOI:10.16707/j.cnki.fjpc.2024.04.010
神经网络在老照片修复中的应用
The Application Of Neural Network In The Restoration Of Old Photos
摘要
Abstract
With the popularity of electronic cameras and scanners,old photos can be digitized and then manually repaired.When repairing photos,manually extracting photo features is both laborious and time-consuming,making it difficult to meet the restoration needs of a large number of old photos.To address this issue,this article proposes an attention based method for repairing old photos.This method not only learns the mapping of specific tasks through the powerful representation ability of neural networks,but also introduces attention mechanisms and knowledge distillation into the framework of old photo restoration networks.The experimental results show that the proposed method can achieve realistic and natural restoration results in repairing texture structure areas of images,significantly improving the visual effect of severely degraded old photos.关键词
照片/图像修复/机器视觉/神经网络Key words
Photos/Image Restoration/Machine Vision/Neural Network分类
信息技术与安全科学引用本文复制引用
杨庆庆,卓树峰,魏丽芬,陈岳林,郑静宜..神经网络在老照片修复中的应用[J].福建电脑,2024,40(4):45-48,4.基金项目
本文得到福建信息职业技术学院校级课题(No.Y22106)资助. （No.Y22106）
