基于PSVM的乳腺X照片分类器设计

纹理是图像中反复出现的局部模式和它们的排列规则。一幅图像的灰度共生矩阵反映了图像灰度关于方向、相邻间隔变化幅度的综合信息，它是分析图像纹理的基础。文中提出了基于灰度共生矩阵的乳腺X照片纹理特征提取方法。同时，利用比BP神经网络精度更高，比传统的SVM算法效率更高、更易于实现的PSVM算法作为分类器的训练方法。     

Optimized Gabor features for mass classification in mammography

Gabor filter bank has been successfully used for false positive reduction problem and the discrimination of benign and malignant masses in breast cancer detection. However, a generic Gabor filter bank is not adapted to multi-orientation and multi-scale texture micro-patterns present in the regions of interest (ROIs) of mammograms. There are two main optimization concerns: how many filters should be in a Gabor filter band and what should be their parameters. Addressing these issues, this work focuses on finding optimizing Gabor filter banks based on an incremental clustering algorithm and Particle Swarm Optimization (PSO). We employ an SVM with Gaussian kernel as a fitness function for PSO. The effect of optimized Gabor filter bank was evaluated on 1024 ROIs extracted from a Digital Database for Screening Mammography (DDSM) using four performance measures (i.e., accuracy, area under ROC curve, sensitivity and specificity) for the above mentioned mass classification problems. The results show that the proposed method enhances the performance and reduces the computational cost. Moreover, the Wilcoxon signed rank test over the significance level of 0.05 reveals that the performance difference between the optimized Gabor filter bank and non-optimized Gabor filter bank is statistically significant.     

Development and validation of an algorithm for registration of serial 3D MR breast data sets

We report on the development of an algorithm to improve the registration of serial 3D MR breast images using combined global translation and rotation with locally varying parameters as geometric transformations. Several phantom and volunteer data sets were acquired and registered using mutual information as a similarity measure of the matching process. After applying a global translation by using a rigid matcher, optimum horizontal and vertical rotation angles were determined. In case of the phantom measurements, angle optimization was performed for each slice of the 3D data set of the phantom, which was deliberately shifted and rotated around different axes. In case of registration of volunteer data, optimum rotation parameters were calculated for preselected equidistant slices of the data set to speed up the calculation time. For slices located between and outside these support slices, the rotation angles were calculated by linear interpolation and extrapolation of the slope of the regression determined by the optimized angles of the support slices. The algorithm improves the registration of serial 3D MR data sets and represents a compromise between a rigid and an elastic 3D matching procedure.     

基于Contourlet变换的医学图像增强

针对传统算法中存在噪声过增强的问题,给出了一种基于Contourlet变换的乳腺X线照片去噪增强算法。Contourlet变换作为一种多尺度几何分析方法,是一种具有多分辨的、局部化和方向化性质的图像表示方法。以此为基础,算法对图像分解后的Contourlet系数进行Stein阈值去噪,然后对不同子带上的各分解系数用非线性增益函数进行不同程度的增强。实验表明,该算法在去除噪声的同时有效突出了乳腺X线照片中的细微特征,有利于小乳腺癌的诊断。最后,文中还通过客观量化指标比较了不同增强算法的效果。     

基于抗混叠轮廓波的乳腺图像降噪与增强方法

乳腺X射线成像是乳腺疾病早期检测的有效手段.然而典型的乳腺X射线图像往往对比度低,噪声污染严重,本文提出一种新颖的基于抗混叠轮廓波变换的乳腺图像降噪及增强方案.首先分析了原始轮廓波变换的频谱混叠问题,设计出一种能稀疏表示图像边界及纹理信息,同时能抑制混叠影响的抗混叠轮廓波变换;在此基础上,分别采用高斯分布与广义拉普拉斯分布来刻划噪声相关及信号相关的变换系数,实现阈值萎缩降噪;接着对处理后的系数进行非线性增强,达到增强乳腺图像中细节信息的效果.实验结果表明,本文方法能有效提高乳腺图像的质量,在计算机辅助乳腺诊断方面有较高应用价值.     

基于注意力机制的乳腺X线摄影分类方法

环境的日益恶化导致癌症的发病率不断升高,2018年全球乳腺癌的发病率在所有癌症中已经位居首位。乳腺X线摄影价格实惠且易于操作,目前被认作是最好的乳腺癌筛查方法,也是早期发现乳腺癌最有效的方法。针对乳腺X线摄影不容易分辨、特征不明显等特点,提出了基于RNN+CNN的注意力记忆网络对其进行分类。注意力记忆网络包含注意力记忆模块和卷积残差模块。注意力记忆模块中,注意力模块提取乳腺X线摄影的特征,记忆模块在RNN网络加入注意力权重来模拟人对所提取关键信息的重点突出;卷积残差模块使用CNN对图像进行分类。该方法创新之处在于：提出注意力记忆网络用于乳腺X线摄影图像分类;所设计网络在RNN+CNN结构上引入注意力权重,提取图像关键信息以增强特征描述。在乳腺X线摄影INbreast数据集上的实验结果显示,注意力记忆网络的运行时间比预训练的Inceptionv2、ResNet50、VGG16网络少50%以上,同时达到更高的分类准确率。     

The effects of message framing and ethnic targeting on mammography use among low-income women.

The authors examined the effects that differently framed and targeted health messages have on persuading low-income women to obtain screening mammograms. The authors recruited 752 women over 40 years of age from community health clinics and public housing developments and assigned the women randomly to view videos that were either gain or loss framed and either targeted specifically to their ethnic groups or multicultural. Loss-framed, multicultural messages were most persuasive. The advantage of loss-framed, multicultural messages was especially apparent for Anglo women and Latinas but not for African American women. These effects were stronger after 6 months than after 12 months. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

人型网络乳腺钼靶影像分类方法研究

乳腺癌是女性最常见的恶性肿瘤之一,严重威胁患者健康,因此乳腺钼靶图像多分类对临床诊断乳腺癌具有十分重要的作用。传统卷积神经网络直接采用高级特征对乳腺钼靶图像进行多分类研究,此方法准确率不高。为了进一步提高分类准确率,构建人型网络模型进行分类。此结构通过堆叠的卷积层以及最大池化层来进行图片的低级特征进行提取,通过堆叠的卷积层以及上池化层将特征逐步返回到图片形式的特征图,通过堆叠的卷积层以及最大池化层再次提取到更高级的特征并与之前的低级特征进行级联,将级联的特征经过全局最大池化层进行池化并得到最终分类。在中山大学肿瘤防治中心的1 824幅乳腺钼靶图像做仿真实验,实验结果表明,该方法的准确率达到了74.54%,优于现有相关网络模型。     

乳腺钼靶X线肿块检测及分割方法

乳腺癌临床中最常见的病理征象是可疑的钙化点及肿块,钙化点的检测技术已经相对成熟,而对肿块进行检测及分割仍是众多学者研究的热点。首先简要介绍了乳腺癌的研究意义、现状及常用的乳腺X线图像库;其次针对目前乳腺肿块的热点研究,重点综述了最新的常用于乳腺肿块检测及分割的方法,对这些方法进行归类,并给出不同方法的优缺点;最后总结并讨论了乳腺X线图像中肿块检测及分割的发展趋势。     

乳腺癌筛查中乳腺X射线摄影两种标准体位及乳腺压迫厚度与平均腺体剂量的关系北大核心CSCD

为了使乳腺X射线摄影在乳腺癌筛查中的应用达到最优化,本文回顾性分析了在我院行乳腺X射线摄影和乳腺彩色多普勒超声检查,并取得病理结果的患者共132人、139个病灶,比较乳腺头尾位(craniocaudal view,CC)、内外侧斜位(medial lateral oblique view,MLO)、乳腺压迫厚度、乳腺密度与被检者接受的平均腺体剂量(average glandular dose,AGD)的关系及两种体位上病灶检出情况。结果显示:(1)乳腺压迫厚度及乳腺密度都是AGD的独立影响因素,并呈正相关,且乳腺压迫厚度对AGD的影响比乳腺密度大;(2)曝光参数管电压及管电流随着乳腺压迫厚度的增加有增加趋势,致使AGD增大;(3)同一乳腺CC及MLO的压迫厚度及AGD未发现特定规律,但均数±标准差CC位(2.49±0.84)>MLO位(2.27±0.81),且MLO比CC更易显示病灶。只拍摄乳腺MLO联合乳腺彩色多普勒超声检查具有很大优势,不仅保证了诊断需要,更降低了被检者的AGD值,且被检者乳腺压迫厚度及乳腺密度越大,AGD降低越多。