四川大学学报（自然科学版）2024，Vol.61Issue(5)：41-51,11.DOI:10.19907/j.0490-6756.2024.052002
基于分数阶全变分和扩散模型的图像去模糊方法
Image deblurring based on fractional-order total variation and diffusion models
摘要
Abstract
Image deblurring is a significant area of study in the field of digital image processing.In various practical applications,due to relative motion between imaging devices and objects,the resulting blurring will degrade the image quality and visual effect.This paper proposes an image deblurring method combining Fractional-order Total Variation(FTV)loss function and The enhancement of image structure information through probability modeling by DDPM,followed by the incorporation of the FTV loss function as a regular-ization term,facilitates the restoration of image details.Compared with traditional image deblurring methods,this method can restore more image detail while maintaining the overall image clarity.The experimental re-sults have verified that the method has the ability to restore images affected by motion blur.The significant su-periority provides a new direction for the further development of image deblurring.关键词
图像去模糊/分数阶微积分/分数阶全变分/去噪扩散概率模型Key words
Image debluring/Fractional calculus/Fractional-order total variation/Denoising diffusion proba-bilistic models分类
信息技术与安全科学引用本文复制引用
黄浩,蒲亦非..基于分数阶全变分和扩散模型的图像去模糊方法[J].四川大学学报（自然科学版）,2024,61(5):41-51,11.基金项目
国家自然科学基金面上项目(62171303) （62171303）
中国兵器装备集团(成都)火控技术中心项目(非密)(HK20-03) （成都）
国家重点研发项目(2018YFC0830300) （2018YFC0830300）
