摘 要
超声成像技术因其无创性、实时性和低成本等优势，在医学诊断领域得到了广泛应用，但其图像质量受噪声干扰、对比度不足及边缘模糊等问题的限制，影响了临床应用效果。为解决上述问题，本研究旨在探索一种高效的超声图像增强方法，以提升图像的视觉清晰度和诊断价值。为此，提出了一种基于深度学习与传统图像处理相结合的混合增强框架，该框架首先利用自适应滤波算法降低图像噪声，随后通过卷积神经网络（CNN）提取图像特征并优化对比度分布，最后采用非线性拉伸技术强化边缘细节。实验结果表明，所提方法在信噪比、结构相似性指数（SSIM）和边缘保持能力等方面均优于传统增强算法和单一深度学习模型。具体而言，与经典直方图均衡化方法相比。
关键词：超声图像增强 深度学习 卷积神经网络
Abstract
Ultrasound imaging technology has been widely used in the field of medical diagnosis due to its advantages of non-invasive, real-time and low cost, but its image quality is limited by noise interference, insufficient contrast and edge blur, which affects the clinical application effect. To address the above issues, this study aims to explore an efficient ultrasound image enhancement method to enhance the visual clarity and diagnostic value of images. To this end, a hybrid enhancement fr amework based on deep learning and traditional image processing is proposed. This fr amework first uses adaptive filtering algorithm to reduce image noise, then extracts image features through convolutional neural network (CNN) and optimizes contrast distribution, and finally non-linear stretching technology is used to strengthen the edge details. The experimental results show that the proposed method is better than the traditional enhancement algorithm and single deep learning model in signal to noise ratio, structural similarity index (SSIM) and edge retention ability. Specifically, as compared to the classical histogram equalization approach.
Keyword:Ultrasound Image Enhancement Deep Learning Convolutional Neural Network
目 录
引言 1
1超声成像技术基础研究 1
1.1超声成像原理概述 1
1.2成像质量的关键因素分析 2
1.3图像增强的技术需求 2
1.4当前图像增强方法的局限性 2
2图像增强算法理论分析 3
2.1常用图像增强算法分类 3
2.2空域增强算法的研究进展 3
2.3频域增强算法的应用现状 4
2.4深度学习在图像增强中的作用 4
3超声图像增强关键技术研究 5
3.1边缘检测与增强技术 5
3.2对比度优化方法探讨 5
3.3噪声抑制与清晰度提升 6
3.4多尺度分析在增强中的应用 6
4实验验证与性能评估 6
4.1数据集构建与实验设计 6
4.2不同算法的对比实验 7
4.3性能指标体系的建立 7
4.4实验结果分析与讨论 8
结论 8
参考文献 10
致谢 11
