[1]郭红建,庄名驹,李嘉豪.一种用于图像超分辨率重建的稠密残差高效网络[J].南京师大学报(自然科学版),2025,48(06):111-120.[doi:10.3969/j.issn.1001-4616.2025.06.012]
 Guo Hongjian,Zhuang Mingju,Li Jiahao.An Efficient Dense Residual Network for Image Super-Resolution Reconstruction[J].Journal of Nanjing Normal University(Natural Science Edition),2025,48(06):111-120.[doi:10.3969/j.issn.1001-4616.2025.06.012]
点击复制

一种用于图像超分辨率重建的稠密残差高效网络()

《南京师大学报(自然科学版)》[ISSN:1001-4616/CN:32-1239/N]

卷:
48
期数:
2025年06期
页码:
111-120
栏目:
计算机科学与技术
出版日期:
2025-12-20

文章信息/Info

Title:
An Efficient Dense Residual Network for Image Super-Resolution Reconstruction
文章编号:
1001-4616(2025)06-0111-10
作者:
郭红建庄名驹李嘉豪
(南京审计大学计算机学院,江苏 南京 211815)
Author(s):
Guo HongjianZhuang MingjuLi Jiahao
(School of Computer Science,Nanjing Audit University,Nanjing 211815,China)
关键词:
深度学习超分辨率残差
Keywords:
deep learningsuper-resolutionresidual
分类号:
TP391
DOI:
10.3969/j.issn.1001-4616.2025.06.012
文献标志码:
A
摘要:
针对超分辨率模型效率低下的问题,本研究深入剖析了其核心结构并提出了一种基于稠密残差的高效网络架构. 该架构在减少模型参数的同时,实现了特征的高效整合,提升了模型效能与推理速度. 通过引入经过优化的残差蒸馏模块,进一步降低了模型的参数规模和计算复杂度,同时摒弃了导致运行效率下降的残差连接方式. 为了增强模型对图像高频细节的捕捉能力,研究还引入了对比损失函数. 最后,通过采用多阶段热启动训练策略,模型的性能得到了提升,实现了更高效、更精准的超分辨率重建.
Abstract:
To address the issue of low efficiency in super-resolution models,we analyze their core structure and propose an efficient network architecture based on dense residuals. This architecture achieves efficient integration of features while reducing model parameters,improving model performance and inference speed. By introducing an optimized residual distillation module,the parameter size and computational complexity of the model have been further reduced,while discarding the residual connection method that leads to decreased operational efficiency. In order to enhance the model's ability to capture high-frequency details in images,the study also introduces a contrastive loss function. Finally,by adopting a multi-stage hot start training strategy,the performance of the model is improved,achieving more efficient and accurate super-resolution reconstruction.

参考文献/References:

[1]钟梦圆,姜麟. 超分辨率图像重建算法综述[J]. 计算机科学与探索,2022,16(05):972-990.
[2]SHI W,CABALLERO J,HUSZÁR F,et al. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Las,Vegas,NV,USA,2016:1874-1883.
[3]周登文,李文斌,李金新,等. 一种轻量级的多尺度通道注意图像超分辨率重建网络[J]. 电子学报,2022,50(10):2336-2346.
[4]LIU J,TANG J,WU G. Residual feature distillation network for lightweight image super-resolution[C]//ECCV Workshops. Glasgow,UK,2020.
[5]何鹏浩,余映,徐超越. 基于动态金字塔和子空间注意力的图像超分辨率重建网络[J]. 计算机科学,2022,49(S2):423-430.
[6]HUANG G,LIU Z,WEINBERGER K Q. Densely connected convolutional networks[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Honolulu,HI,USA,2016:2261-2269.
[7]ZHANG Y,TIAN Y,KONG Y,et al. Residual dense network for image super-resolution[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,UT,USA,2018:2472-2481.
[8]HARIS M,SHAKHNAROVICH G,UKITA N. Deep back-projection networks for super-resolution[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,UT,USA,2018:1664-1673.
[9]LIU J,ZHANG W,TANG Y,et al. Residual feature aggregation network for image super-resolution[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Seattle,WA,USA:2020:2356-2365.
[10]HUI Z,WANG X,GAO X. Fast and accurate single image super-resolution via information distillation network[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,UT,USA,2018:723-731.
[11]LI J,FANG F,MEI K,et al. Multi-scale residual network for image super-resolution[C]//European Conference on Computer Vision. Munich,Germany,2018.
[12]LIM B,SON S,KIM H,et al. Enhanced deep residual networks for single image super-resolution[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition workshops(CVPRW). Honolulu,HI,USA,2017:1132-1140.
[13]叶汉民,李志波,程小辉,等. 基于注意力机制的稀疏化剪枝方法[J]. 计算机工程与设计,2023,44(12):3642-3648. DOI:10.16208/j.issn1000-7024.2023.12.017.
[14]SUMIT K,KUMAR R S. Robust denoising technique for ultrasound images by splicing of low rank filter and principal component analysis[J]. Biomedical research,2018,29(18):3444-3455.
[15]WEI L,MA Z,YANG C,et al. Advances in the neural network quantization:a comprehensive review[J]. Applied sciences,2024,14(17):7445-7445.
[16]杨木润,曹润柘,杜权,等. 神经网络结构搜索前沿综述[J]. 中文信息学报,2023,37(10):1-15.
[17]顿家乐,王骏,彭汉琛,等. 面向自闭症辅助诊断的知识蒸馏混合域适应方法[J]. 智能系统学报,1-10[2025-02-22]. http://kns.cnki.net/kcms/detail/23.1538.TP.20250126.1344.004.html.
[18]JUNWEI L,MUJIAN Y,XIAOLIN Y,et al. Binary image steganalysis based on symmetrical local residual patterns[J]. Chinese journal of electronics,2022,31(4):752-763.
[19]ZHANG Q,MIN B,HANG Y,et al. A full-scale lung image segmentation algorithm based on hybrid skip connection and attention mechanism[J]. Scientific reports,2024,14(1):23233-23233.
[20]IBARRA D H O,SARGSYAN K,NAJM N H. Surrogate construction via weight parameterization of residual neural networks[J]. Computer methods in applied mechanics and engineering,2025,433(PA):117468-117468.
[21]YE H,HUANG X,ZHU H,et al. An enhanced network with parallel graph node diffusion and node similarity contrastive loss for hyperspectral image classification[J]. Digital signal processing,2025,158:104965.
[22]张建华,孔繁涛,吴建寨,等. 基于改进VGG卷积神经网络的棉花病害识别模型[J]. 中国农业大学学报,2018,23(11):161-171.
[23]KONG F,LI M,LIU S,et al. Residual Local Feature Network for Efficient Super-Resolution[M]. arXiv,2022.
[24]ZHANG B,ZHANG Y,WANG B,et al. Denoising swin transformer and perceptual peak signal-to-noise ratio for low-dose ct image denoising[J]. Measurement,2024,227:114303.
[25]ILLYA B,MARCO B,RAIMONDO S,et al. Structural similarity index(ssim)revisited:a data-driven approach[J]. Expert systems with applications,2022,189:116087.
[26]SIAN C,YAJUAN Z,RUI W. RETRACTED:An optimal framework for natural english translate processing:on the application of artificial neural network and adam optimization algorithm[J]. Journal of intelligent & fuzzy systems,2024,47(5-6,Supplement 1):205-219.

相似文献/References:

[1]郑德鹏,杜吉祥,翟传敏.基于深度学习MPCANet的年龄估计[J].南京师大学报(自然科学版),2017,40(01):20.[doi:10.3969/j.issn.1001-4616.2017.01.004]
 Zheng Depeng,Du Jixiang,Zhai Chuanmin.Age Estimation Based on Deep Learning MPCANet[J].Journal of Nanjing Normal University(Natural Science Edition),2017,40(06):20.[doi:10.3969/j.issn.1001-4616.2017.01.004]
[2]朱 繁,王洪元,张 继.基于深度学习的行人重识别研究综述[J].南京师大学报(自然科学版),2018,41(04):93.[doi:10.3969/j.issn.1001-4616.2018.04.015]
 Zhu Fan,Wang Hongyuan,Zhang Ji.A Survey of Person Re-identification Based on Deep Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2018,41(06):93.[doi:10.3969/j.issn.1001-4616.2018.04.015]
[3]孙茹君,张鲁飞.基于动态指导的深度学习模型稀疏化执行方法[J].南京师大学报(自然科学版),2019,42(03):11.[doi:10.3969/j.issn.1001-4616.2019.03.002]
 Sun Rujun,Zhang Lufei.Dynamic Sparse Method for Deep Learning Execution[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(06):11.[doi:10.3969/j.issn.1001-4616.2019.03.002]
[4]赵文芳,林润生,唐 伟,等.基于深度学习的PM2.5短期预测模型[J].南京师大学报(自然科学版),2019,42(03):32.[doi:10.3969/j.issn.1001-4616.2019.03.005]
 Zhao Wenfang,Lin Runsheng,Tang Wei,et al.Forecasting Model of Short-Term PM2.5 ConcentrationBased on Deep Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(06):32.[doi:10.3969/j.issn.1001-4616.2019.03.005]
[5]张新峰,闫昆鹏,赵 珣.基于双向LSTM的手写文字识别技术研究[J].南京师大学报(自然科学版),2019,42(03):58.[doi:10.3969/j.issn.1001-4616.2019.03.008]
 Zhang Xinfeng,Yan Kunpeng,Zhao Xun.Handwriting Chinese Text Recognition Using BiLSTM Network[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(06):58.[doi:10.3969/j.issn.1001-4616.2019.03.008]
[6]贾玉福,胡胜红,刘文平,等.使用条件生成对抗网络的自然图像增强方法[J].南京师大学报(自然科学版),2019,42(03):88.[doi:10.3969/j.issn.1001-4616.2019.03.012]
 Jia Yufu,Hu Shenghong,Liu Wenping,et al.Wild Image Enhancement with Conditional Generative Adversarial Network[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(06):88.[doi:10.3969/j.issn.1001-4616.2019.03.012]
[7]汤 凯,何 庆,赵 群,等.基于改进的深度残差网络的图像识别[J].南京师大学报(自然科学版),2019,42(03):115.[doi:10.3969/j.issn.1001-4616.2019.03.015]
 Tang Kai,He Qing,Zhao Qun,et al.Image Recognition Based on Improved Deep Neural Network[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(06):115.[doi:10.3969/j.issn.1001-4616.2019.03.015]
[8]汪 晨,张辉辉,乐继旺,等.基于深度学习和遥感影像的松材线虫病疫松树目标检测[J].南京师大学报(自然科学版),2021,44(03):84.[doi:10.3969/j.issn.1001-4616.2021.03.013]
 Wang Chen,Zhang Huihui,Le Jiwang,et al.Object Detection to the Pine Trees Affected by Pine Wilt Diseasein Remote Sensing Images Using Deep Learning[J].Journal of Nanjing Normal University(Natural Science Edition),2021,44(06):84.[doi:10.3969/j.issn.1001-4616.2021.03.013]
[9]韩 悦,张永寿,郭依廷,等.乳腺癌腋窝淋巴结超声图像分割算法研究[J].南京师大学报(自然科学版),2021,44(04):122.[doi:10.3969/j.issn.1001-4616.2021.04.016]
 Han Yue,Zhang Yongshou,Guo Yiting,et al.Research on Ultrasound Image Segmentation Algorithm forAxillary Lymph Node with Breast Cancer[J].Journal of Nanjing Normal University(Natural Science Edition),2021,44(06):122.[doi:10.3969/j.issn.1001-4616.2021.04.016]
[10]闫靖昆,黄毓贤,秦伟森,等.棉田复杂背景下棉花黄萎病病斑分割算法研究[J].南京师大学报(自然科学版),2021,44(04):127.[doi:10.3969/j.issn.1001-4616.2021.04.017]
 Yan Jingkun,Huang Yuxian,Qin Weisen,et al.Study on Segmentation Algorithm of Cotton Verticillium WiltDisease Spot in Cotton Field Under Complex Background[J].Journal of Nanjing Normal University(Natural Science Edition),2021,44(06):127.[doi:10.3969/j.issn.1001-4616.2021.04.017]

备注/Memo

备注/Memo:
收稿日期:2025-09-02.
基金项目:国家自然科学基金面上资助项目(62375133)、江苏省高校自然科学研究资助项目(25KJA520006)、江苏省产学研合作资助项目(BY20230589)、2024年度江苏省社科应用研究精品工程课题资助项目(24SYB-121).
通讯作者:郭红建,博士,副教授,研究方向:数据挖掘,大数据审计. E-mail:g_coolman@163.com
更新日期/Last Update: 2025-12-20