本文为美国西卡罗莱纳大学（作者：HongdaShen）的硕士论文，共38页。

据美国癌症协会报道，乳腺癌是女性所有癌症中第二大死因。另据报道，早期发现乳腺癌可以通过允许更广泛的治疗选择来提高生存率。乳腺X光摄影术被认为是一种有效的工具，可以帮助放射科医生早期发现恶性乳腺癌。图像增强技术可以通过增强关键特征的细节，如微钙化点的形状来提高乳腺X光图像的质量。

本论文提出了一种新的乳房X光图像增强方法。该方法采用三层拉普拉斯金字塔（LP）方案，用压缩盒滤波器（SBF）代替传统的低通滤波。将先前提出的非线性局部增强技术应用于Laplacian金字塔产生的差分图像，以对比度增强乳腺图像的结构细节。通过将增强后的差分图像加入到原始的SBF滤波图像中，重建增强后的乳腺图像。实验和定量结果为本文提出的图像增强方法对乳腺图像的鲁棒性提供了经验证据。

Breast cancer has been reported by AmericanCancer Society as the second leading cause of death among all the cancers ofwomen. It is also reported that the early detection of breast cancer canimprove survival rate by allowing a wider range of treatment options.Mammography is believed to be an effective tool to help radiologists to detectthe malignant breast cancer at the early stage. Image enhancement techniquescan improve the quality of mammogram images with enhancing the details of keyfeatures, like the shape of microcalcifications. This thesis proposed a novelmethod to enhance mammogram images. The proposed method uses a three levelLaplacian Pyramid (LP) scheme that applies the Squeeze Box Filter (SBF) insteadof conventional low pass filtering. A previously proposed nonlinear localenhancement technique is applied to the difference image produced in theLaplacian Pyramid to contrast enhance the structural details of mammogramimages. The enhanced mammogram image is reconstructed by adding all theenhanced difference images to the origianl SBF filtered image. Experimentationand quantitative results reported in this thesis provide empirical evidence onthe robustness of the proposed image enhancement method on mammographic images.

1.      引言

2. 文献回顾

3. 研究方法

4. 实验结果与讨论

5. 结论与展望

更多精彩文章请关注公众号：