摘要
由于数字媒体信息共享系统中部分图片存在失真、噪声大、边缘丢失等问题,为此,提出一种图片质量增强方法.构建由INM模块和URM模块组成的双网络模型结构,对低照度图像经过特征提取、融合、重建后,得到正常照度图像;在损失函数的作用下,尽可能多地保留图像边缘细节信息,同时结合对比度增强算法,实现对图像整体质量的增强.经实验验证,创建的方法可在保证图片失真程度最小的前提下,完成高质量的图像增强.
Because some pictures in digital media information-sharing system have some problems such as distortion,high noise,edge loss and so on,a method of image quality enhancement is proposed.The dual net-work model structure composed of the INM module and URM module is constructed.After feature extrac-tion,fusion and reconstruction of low illumination images,normal illumination images are obtained;Under the effect of the loss function,the image edge details are retained as much as possible,and the overall image quality is enhanced by combining the contrast enhancement algorithm.Results show that the proposed meth-od can achieve high quality image enhancement with minimum distortion.
作者
王艳
苏怡航
WANG Yan;SU Yi-hang(College of Art,Anhui Jianzhu University,Hefei 230041,Anhui,China)
出处
《喀什大学学报》
2023年第3期77-80,共4页
Journal of Kashi University
基金
2021年度安徽省教育厅科学研究项目“新媒体语境下安徽非物质文化遗产数字化发展空间及传播路径研究”(SK2021JD12).
关键词
图片质量
损失函数
URM模块
对比度增强
特征提取
双网络模型结构
image quality
loss function
URM module
contrast enhancement
feature extraction
dual network model structure
