本科生毕业论文

发布时间：2025年09月27日作者:aiycxz.cn

题目基于深度学习的图像超分辨率重建学院信息科学与工程学院专业计算机科学与技术毕业届别 2022届姓名李明导师马忠彧职称副教授# 西北师范大学教务处制## 目录**摘要** ...... I ABSTRACT ...... II **1 绪论** ...... 1 1.1 研究背景及意义 ...... 1 1.2 国内外研究现状 ...... 2 1.3 论文主要工作 ...... 3 1.4 论文结构安排 ...... 4 **2 图像超分辨率重建相关技术** ...... 5 2.1 图像超分辨率重建基本概念 ...... 5 2.2 图像超分辨率重建评价指标 ...... 6 2.3 基于插值的图像超分辨率重建 ...... 7 2.4 基于重建的图像超分辨率重建 ...... 8 2.5 基于深度学习的图像超分辨率重建 ...... 9 2.6 本章小结 ...... 10 **3 基于残差密集网络的图像超分辨率重建** ...... 11 3.1 残差密集网络 ...... 11 3.2 残差密集网络结构 ...... 12 3.3 残差密集模块 ...... 13 3.4 实验设置 ...... 14 3.5 实验结果与分析 ...... 15 3.6 本章小结 ...... 16 **4 基于注意力机制的图像超分辨率重建** ...... 17 4.1 注意力机制 ...... 17 4.2 残差注意力网络 ...... 18 4.3 残差注意力网络结构 ...... 19 4.4 残差注意力模块 ...... 20 4.5 实验设置 ...... 21 4.6 实验结果与分析 ...... 22 4.7 本章小结 ...... 23 **5 总结与展望** ...... 24 5.1 总结 ...... 24 5.2 展望 ...... 25 **参考文献** ...... 26 **致谢** ...... 28 基于深度学习的图像超分辨率重建摘要图像超分辨率重建技术是指将给定的低分辨率图像重建为高分辨率图像的技术，在医学影像、遥感图像、视频监控等领域有着广泛的应用。近年来，随着深度学习的发展，基于深度学习的图像超分辨率重建方法取得了显著的成果。本文主要研究基于深度学习的图像超分辨率重建方法，并针对现有方法中存在的问题进行改进。首先，本文介绍了图像超分辨率重建的基本概念、评价指标以及传统的图像超分辨率重建方法，包括基于插值的方法和基于重建的方法。然后，重点介绍了基于深度学习的图像超分辨率重建方法，包括卷积神经网络、残差网络、生成对抗网络等。其次，本文提出了一种基于残差密集网络的图像超分辨率重建方法。该方法通过引入残差密集模块，增强了网络的特征提取能力，提高了重建图像的质量。实验结果表明，该方法在峰值信噪比和结构相似性指标上均优于传统的超分辨率重建方法。再次，本文提出了一种基于注意力机制的图像超分辨率重建方法。该方法通过引入注意力机制，使网络能够自适应地关注图像中的重要区域，从而提高了重建图像的细节保留能力。实验结果表明，该方法在视觉效果和客观评价指标上均优于现有的超分辨率重建方法。最后，本文对提出的两种方法进行了总结，并展望了未来的研究方向。关键词：图像超分辨率重建；深度学习；残差密集网络；注意力机制1# Image Super-Resolution Reconstruction Based on Deep Learning## ABSTRACTImage super-resolution reconstruction technology refers to the technology of reconstructing a given low-resolution image into a high-resolution image, which has a wide range of applications in medical imaging, remote sensing images, video surveillance and other fields. In recent years, with the development of deep learning, image super-resolution reconstruction methods based on deep learning have achieved significant results. This paper mainly studies the image super-resolution reconstruction method based on deep learning, and improves the existing problems in the existing methods.First, this paper introduces the basic concepts of image super-resolution reconstruction, evaluation indicators, and traditional image super-resolution reconstruction methods, including interpolation-based methods and reconstruction-based methods. Then, it focuses on the image super-resolution reconstruction method based on deep learning, including convolutional neural networks, residual networks, generative adversarial networks, etc.Secondly, this paper proposes an image super-resolution reconstruction method based on residual dense network. By introducing residual dense modules, this method enhances the feature extraction capability of the network and improves the quality of the reconstructed image. Experimental results show that this method is superior to traditional super-resolution reconstruction methods in terms of peak signal-to-noise ratio and structural similarity index.Thirdly, this paper proposes an image super-resolution reconstruction method based on attention mechanism. By introducing the attention mechanism, this method enables the network to adaptively focus on important areas in the image, thereby improving