基于不同超声成像的甲状腺结节良恶性判别

为实现更为准确的甲状腺结节良恶性超声图像诊断,避免不必要的穿刺或活检手术,提出了一种基于卷积神经网络（CNN）的常规超声成像和超声弹性成像的特征结合方法,提高了甲状腺结节良恶性分类准确率。首先,卷积网络模型在大规模自然图像数据集上完成预训练,并通过迁移学习的方式将特征参数迁移到超声图像域用以生成深度特征并处理小样本。然后,结合常规超声成像和超声弹性成像的深度特征图形成混合特征空间。最后,在混合特征空间上完成分类任务,实现了一个端到端的卷积网络模型。在1156幅图像上进行实验,所提方法的准确率为0.924,高于其他单一数据源的方法。实验结果表明,浅层卷积共享图像的边缘纹理特征,高层卷积的抽象特征与具体的分类任务相关,使用迁移学习的方法可以解决数据样本不足的问题;同时,弹性超声影像可以对甲状腺结节的病灶硬度进行客观的量化,结合常规超声的纹理轮廓特征,二者融合的混合特征可以更全面地描述不同病灶之间的差异。所提方法可以高效准确地对甲状腺结节进行良恶性分类,减轻患者痛苦,给医生提供更为准确的辅助诊断信息。     

基于多尺度融合的甲状腺结节图像特征提取

甲状腺结节是一种常见的多发病,超声技术是该疾病首选的检查方法。在超声图像中提取区分甲状腺结节良恶性的纹理特征并进行判别具有广阔的临床应用前景。双树复小波变换（Dual-tree complex wavelet transform,DT-CWT）和Gabor小波是纹理特征提取的常用方法。本文提出一种基于多尺度的DT-CWT和Gabor特征融合的甲状腺结节识别方法。该方法首先通过高斯金字塔将甲状腺超声图像分解到多尺度空间,然后提取图像的DT-CWT和Gabor的多尺度特征,最后实现特征融合。通过应用支持向量机（Support vector machine,SVM）分类器实现分类,验证特征提取方法的有效性。实验结果表明,本文提出的方法能达到较高的识别率。     

融合深度网络和浅层纹理特征的甲状腺结节癌变超声图像诊断

目的 在甲状腺结节图像中对甲状腺结节进行良恶性分析,对于甲状腺癌的早期诊断有着重要的意义。随着医疗影像学的发展,大部分的早期甲状腺结节可以在超声图像中准确地检测出来,但对于结节的性质仍然缺乏准确的判断。因此,为实现更为准确的早期甲状腺结节良恶性超声图像诊断,避免不必要的针刺或其他病理活检手术、减轻病患生理痛苦和心理压力及其医疗费用,提出一种基于深度网络和浅层纹理特征融合的甲状腺结节良恶性分类新算法。方法 本文提出的甲状腺结节分类算法由4步组成。首先对超声图像进行尺度配准、人工标记以及图像复原去除以增强图像质量。然后,对增强的图像进行数据扩展,并作为训练集对预训练过的GoogLeNet卷积神经网络进行迁移学习以提取图像中的深度特征。同时,提取图像的旋转不变性局部二值模式（LBP）特征作为图像的纹理特征。最后,将深度特征与图像的纹理特征相融合并输入至代价敏感随机森林分类器中对图像进行良恶性分类。结果 本文方法在标准的甲状腺结节癌变数据集上对甲状腺结节图像取得了正确率99.15%,敏感性99.73%,特异性95.85%以及ROC曲线下面积0.997 0的的好成绩,优于现有的甲状腺结节图像分类方法。结论 实验结果表明,图像的深度特征可以描述医疗超声图像中病灶的整体感官特征,而浅层次纹理特征则可以描述超声图像的边缘、灰度分布等特征,将二者统一的融合特征则可以更为全面地描述图像中病灶区域与非病灶区域之间的差异以及不同病灶性质之间的差异。因此,本文方法可以准确地对甲状腺结节进行分类从而避免不必要手术、减轻病患痛苦和压力。     

A hybrid multi-scale model for thyroid nodule boundary detection on ultrasound images

A hybrid model for thyroid nodule boundary detection on ultrasound images is introduced. The segmentation model combines the advantages of the “á trous” wavelet transform to detect sharp gray-level variations and the efficiency of the Hough transform to discriminate the region of interest within an environment with excessive structural noise. The proposed method comprise three major steps: a wavelet edge detection procedure for speckle reduction and edge map estimation, based on local maxima representation. Subsequently, a multiscale structure model is utilised in order to acquire a contour representation by means of local maxima chaining with similar attributes to form significant structures. Finally, the Hough transform is employed with ‘a priori’ knowledge related to the nodule's shape in order to distinguish the nodule's contour from adjacent structures. The comparative study between our automatic method and manual delineations demonstrated that the boundaries extracted by the hybrid model are closely correlated with that of the physicians. The proposed hybrid method can be of value to thyroid nodules’ shape-based classification and as an educational tool for inexperienced radiologists.     

Automated delineation of thyroid nodules in ultrasound images using spatial neutrosophic clustering and level set

An accurate contour estimation plays a significant role in classification and estimation of shape, size, and position of thyroid nodule. This helps to reduce the number of false positives, improves the accurate detection and efficient diagnosis of thyroid nodules. This paper introduces an automated delineation method that integrates spatial information with neutrosophic clustering and level-sets for accurate and effective segmentation of thyroid nodules in ultrasound images. The proposed delineation method named as Spatial Neutrosophic Distance Regularized Level Set (SNDRLS) is based on Neutrosophic L-Means (NLM) clustering which incorporates spatial information for Level Set evolution. The SNDRLS takes rough estimation of region of interest (ROI) as input provided by Spatial NLM (SNLM) clustering for precise delineation of one or more nodules. The performance of the proposed method is compared with level set, NLM clustering, Active Contour Without Edges (ACWE), Fuzzy C-Means (FCM) clustering and Neutrosophic based Watershed segmentation methods using the same image dataset. To validate the SNDRLS method, the manual demarcations from three expert radiologists are employed as ground truth. The SNDRLS yields the closest boundaries to the ground truth compared to other methods as revealed by six assessment measures (true positive rate is 95.45 ± 3.5%, false positive rate is 7.32 ± 5.3% and overlap is 93.15 ± 5. 2%, mean absolute distance is 1.8 ± 1.4 pixels, Hausdorff distance is 0.7 ± 0.4 pixels and Dice metric is 94.25 ± 4.6%). The experimental results show that the SNDRLS is able to delineate multiple nodules in thyroid ultrasound images accurately and effectively. The proposed method achieves the automated nodule boundary even for low-contrast, blurred, and noisy thyroid ultrasound images without any human intervention. Additionally, the SNDRLS has the ability to determine the controlling parameters adaptively from SNLM clustering.     

基于注意力机制多任务的肺结节癌变风险判断方法

对于CT影像中检测出的肺部结节, 需要自动判断其是否有癌变风险. 不同于大多数现有的研究方法只区分结节良恶性, 本文提出了一个基于注意力机制的多任务学习模型, 将与结节良恶性相关的语义特征属性一并判断输出, 通过判断9个结节特征(对比度、分叶征、毛刺征、球形度、边缘、纹理、钙化程度、大小以及恶性程度)的同时实现内在特征的共享, 以达到提高各子任务性能的目的. 选择视觉转换器(ViT)模型作为多任务共享特征提取层, 整体模型采用动态加权平均方法来对各子任务的Loss函数进行优化. 在LUNA16数据集上的实验表明, 该学习框架可以提升肺结节癌变风险判断的性能, 且同时对其他语义特征的判断也能提升结果的可解释性.     

改进Mask R-CNN的甲状腺结节超声图像分割方法

甲状腺结节超声图像对比度低,斑点噪声严重,且不同病人的甲状腺结节形态差异较大,这给医生准确分割结节带来极大困难。为了精确地从超声图像中分割出甲状腺结节,对原掩膜区域卷积神经网络（mask regionconvolutional neural network,Mask R-CNN）的主干网络进行改进。在原主干网络的残差网络层中加入注意力机制模块来提高模型收敛性,并且在特征金字塔网络中增添一条由下向上的支路,将该支路输出特征图进行融合后,输入至区域推荐网络和感兴趣区域池化层,从而能够在融合多尺度特征的同时平衡特征图信息差异。经过对600幅甲状腺结节超声图像进行测试,改进后Mask R-CNN图像分割的平均Dice系数为0.914 8,平均精确度为0.932 2,平均召回率为0.903 4,平均F1分数为0.917 6。改进算法分割的Dice系数比原Mask R-CNN提升了0.080 6,改进算法可以应用于实际临床医学中自动分割甲状腺结节超声图像。     

A hybrid multi-scale model for thyroid nodule boundary detection on ultrasound images

A hybrid model for thyroid nodule boundary detection on ultrasound images is introduced. The segmentation model combines the advantages of the "á trous" wavelet transform to detect sharp gray-level variations and the efficiency of the Hough transform to discriminate the region of interest within an environment with excessive structural noise. The proposed method comprise three major steps: a wavelet edge detection procedure for speckle reduction and edge map estimation, based on local maxima representation. Subsequently, a multiscale structure model is utilised in order to acquire a contour representation by means of local maxima chaining with similar attributes to form significant structures. Finally, the Hough transform is employed with 'a priori' knowledge related to the nodule's shape in order to distinguish the nodule's contour from adjacent structures. The comparative study between our automatic method and manual delineations demonstrated that the boundaries extracted by the hybrid model are closely correlated with that of the physicians. The proposed hybrid method can be of value to thyroid nodules' shape-based classification and as an educational tool for inexperienced radiologists.     

融合多维度卷积神经网络的肺结节分类方法

针对CT图像肺结节分类任务中分类精度低,假阳性高的问题,提出了一种加权融合多维度卷积神经网络的肺结节分类模型,该模型包含两个子模型：基于二维图像的多尺度密集卷积网络模型,以捕获更宽泛的结节变化特征并促进特征重用;基于三维图像的三维卷积神经网络模型,以充分利用结节空间上下文信息。使用二维和三维CT图像训练子模型,根据子模型分类误差计算其权重,对子模型分类结果进行加权融合,得到最终分类结果。该模型在公共数据集LIDC-IDRI上分类准确率达到94.25%,AUC值达到98%。实验结果表明,加权融合多维度模型可以有效地提升肺结节分类性能。     

改进U型残差网络用于肺结节检测

针对计算机断层扫描（CT）影像中肺结节检测灵敏度较低,且存在大量假阳性的问题,提出一种改进的U型残差网络用于肺结节检测。采取U-net网络的U型结构并利用残差学习方式构建深层次网络,同时引入自校正卷积增加特征的信息提取能力,进行通道间与局部信息增强,有利于检测不同形态的结节;通过引入的通道注意力机制,对特征提取过程中的特征进行重标定,实现自适应学习特征权重,进一步提高检测的准确率;引入DR loss作为该算法的分类损失函数,用于解决数据正负样本失衡问题。在LUNA16数据集对所提算法进行了验证,CPM得分达到0.901,提高了肺结节检测的灵敏度,而且有效降低了检测结果的平均假阳性个数,可有效辅助放射科医师对肺结节进行检测。     

A genetically optimized level set approach to segmentation of thyroid ultrasound images

This paper presents a novel framework for thyroid ultrasound image segmentation that aims to accurately delineate thyroid nodules. This framework, named GA-VBAC incorporates a level set approach named Variable Background Active Contour model (VBAC) that utilizes variable background regions, to reduce the effects of the intensity inhomogeneity in the thyroid ultrasound images. Moreover, a parameter tuning mechanism based on Genetic Algorithms (GA) has been considered to search for the optimal VBAC parameters automatically, without requiring technical skills. Experiments were conducted over a range of ultrasound images displaying thyroid nodules. The results show that the proposed GA-VBAC framework provides an efficient, effective and highly objective system for the delineation of thyroid nodules.     

融合注意力机制和特征金字塔网络的CT图像肺结节检测

目的 针对现有肺结节检测算法存在的因肺部计算机断层扫描（computed tomography,CT）图像肺结节与周边组织复杂性导致结节本身结构差异性不明显的问题,以及特征提取网络多次下采样造成图像分辨率降低进而导致检测结果差、仅使用网络顶层特征图进行预测造成图像空间信息丢失进而导致小结节漏检等问题,提出了一种基于注意力机制和特征金字塔的肺结节检测算法。方法 根据语义与空间特征补偿机制以及卷积神经网络中网络深度所提取特征的信息量不同,在以ResNet为骨干网络的特征提取网络中设计通道—空间注意力机制,尽可能同时获取含有较多上下文语义以及空间位置信息的特征信息。在网络预测部分设计特征金字塔网络,将高维带有丰富语义信息的特征图与低维带有位置信息的特征图融合进行多尺度预测,增强网络对于小结节以及近血管结节等非显著性目标的检测性能。结果 在LUNA16（lung nodule analysis 16）数据集上进行十折交叉验证显示,当平均假阳性个数为25.99时敏感度达到了97.13%,与基准方法相比,敏感度提高了2.53%,平均假阳性降低了28.54,实现了高敏感度低假阳性;在0.125、0.25、0.5、1、2、4、8这7个假阳率点的敏感度平均值为0.854,其中在每个扫描4次和8次假阳性时敏感度分别达到了0.940和0.951,其效果优于主流的结节检测方法。结论 提出的结节检测模型,可以提高对3~10 mm小结节、近血管结节等非显著性目标的检测性能,并具有较低的假阳率。     

基于改进LVQ算法的肺结节识别与分类

目前,肺癌的是发病率最高的肿瘤,若能在早期发现癌变并进行相应治疗,将极大的提高患者的生存率。肺癌的症状在早期表现为肺结节。以提高肺结节检测识别率并进行良恶性分类为目的,提出了一种改进的LVQ分类器算法。首先使用C-V算法对原始图像进行肺实质分割,再使用最优阈值法进行感兴趣区域提取,并进行特征提取和特征归一化。使用多次聚类算法检测肺结节。使用基于改进的LVQ分类器进行肺结节的良恶性进行分类。利用改进后的LVQ分类器在LIDC数据集上进行实验,得到了对良性结节的确诊率为87.3%,对恶性结节的确诊率为80.8%。实验结果表明,改进后的算法在良恶性结节分类上具有较高的确诊率,有助于提高医生的工作效率,实现肺结节的辅助发现。     

ThyroScreen system: high resolution ultrasound thyroid image characterization into benign and malignant classes using novel combination of texture and discrete wavelet transform

Using right equipment and well trained personnel, ultrasound of the neck can detect a large number of non-palpable thyroid nodules. However, this technique often suffers from subjective interpretations and poor accuracy in the differential diagnosis of malignant and benign thyroid lesions. Therefore, we developed an automated identification system based on knowledge representation techniques for characterizing the intra-nodular vascularization of thyroid lesions. Twenty nodules (10 benign and 10 malignant), taken from 3-D high resolution ultrasound (HRUS) images were used for this work. Malignancy was confirmed using fine needle aspiration biopsy and subsequent histological studies. A combination of discrete wavelet transformation (DWT) and texture algorithms were used to extract relevant features from the thyroid images. These features were fed to different configurations of AdaBoost classifier. The performance of these configurations was compared using receiver operating characteristic (ROC) curves. Our results show that the combination of texture features and DWT features presented an accuracy value higher than that reported in the literature. Among the different classifier setups, the perceptron based AdaBoost yielded very good result and the area under the ROC curve was 1 and classification accuracy, sensitivity and specificity were 100%. Finally, we have composed an Integrated Index called thyroid malignancy index (TMI) made up of these DWT and texture features, to facilitate distinguishing and diagnosing benign or malignant nodules using just one index or number. This index would help the clinicians in more quantitative assessment of the thyroid nodules.     

肺结节球表面网格向量化特征分类

目的 基于球谐函数与容斥映射算法向量化球面表面纹理与结节形状用以进行胸部CT图像肺结节良恶性判定。区别于基于深度学习解决肺结节良恶性筛查的方法,目前方法多集中于框架改进而忽略了数据预处理,文中所提方法旨在对球面纹理与结节形状进行向量表达,使其可以输入深度森林进行特征分类训练。方法 首先采用辽宁中医药大学附属医院数据,通过3维重构获得3维肺结节图像。其次使用球谐函数与容斥映射算法在保留空间信息的同时将纹理以网格方式映射到标准球面上。再次使用网格-LBP与映射形变能量分别完成对球面纹理与结节形状信息的构建。最后提出一种基于网格的多粒度扫描方法对深度森林训练框架进行改进,并将向量化后的纹理和形状特征加入到改进的深度森林训练框架中进行实验验证。结果 通过大量的实验结果验证,在准确率（ACC）、特异度（SPE）、敏感度（SEN）和受试者工作特征曲线下的面积（AUC）4个衡量指标下,本文方法具有优于现存先进方法的表现,其中ACC、SPE、SEN和AUC分别达到76.06%、69.46%、88.46%和0.84。结论 基于球谐函数与容斥映射算法可成功地对肺结节表面和形状两个特征进行向量化并训练,不仅考虑了数据预处理,而且通过两个特征对肺结节良恶性检测的准确率要高于传统1个特征检测的结果,同时也为3维模型中特征的提取及向量化提供了一个有效的方法。     

基于局部灰度最大和改进Mahalanobis

CT图像中肺结节检测一直是肺癌CAD系统的关键和难点。提出了一种孤立性肺结节自动检测算法,首先对原始CT图像进行有效、准确的肺实质分割;采用寻找局部灰度最大值方法对ROI进行初始分割;再对分割出的各ROI进行特征提取,利用SVM方法对每个特征进行定量描述,根据SVM单特征分类准确率对Mahalanobis距离进行加权改进,最后采用基于改进的Mahalanobis距离进行肺结节分类。实验结果表明,该算法可以较好地提取出CT图像中的孤立性肺结节,具有较高的灵敏度和较低的漏诊率,可以为医生诊断早期肺癌病灶提供帮助信息。     

Automatic classification for solitary pulmonary nodule in CT image by fractal analysis based on fractional Brownian motion model

Perfusion computed tomography (CT) method has been used to differentiate malignant pulmonary nodules from benign nodules based on the assessment for the change of the CT attenuation value within the pulmonary nodules. Instead of using the change of the CT attenuation value, a set of fractal features based on fractional Brownian motion model is proposed in this paper to automatically distinguish malignant nodules from benign nodules. In a set of 107 CT images from 107 different patients with each image containing a solitary pulmonary nodule, our experimental results obtained from a support vector machine classifier show that the accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and the area under the ROC curve are 83.11%, 90.92%, 71.70%, 80.05%, 87.52%, and 0.8437, respectively, by using the proposed fractal-based feature set. Such a result outperforms the conventional method of using the change of the CT attenuation value as the feature for classification. When combining this conventional method with our proposed fractal-based method, the accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and the area under the ROC curve can be promoted to 88.82%, 93.92%, 82.90%, 87.30%, 90.20%, and 0.9019, respectively. In other words, a high performance of pulmonary nodule classification can be achieved with a single post-contrast CT scan.     

多阶U-Net甲状腺超声图像自动分割方法

甲状腺超声图像广泛应用于甲状腺相关疾病的诊断。针对甲状腺超声图像对比度低、边缘模糊以及散斑噪声严重等问题,提出一种基于多阶U-Net的深度卷积网络模型,用于实现甲状腺腺体与甲状腺结节的自动分割。该模型以U-Net为基本网络框架,通过不断进阶的特征融合,以实现图像边缘的信息提取。同时,在模型中使用了一种多尺度残差卷积模块以进一步提升分割精度。对比实验结果表明,该模型相较于其他方法能够获得更好的分割结果,具有一定的临床应用价值。     

Multiple ROI selection based focal liver lesion classification in ultrasound images

Ultrasound imaging is one of the most widely used imaging modality for the purpose of visualizing the human soft tissues. Especially, liver imaging application is of great importance in the areas of diagnostic ultrasound. In ultrasound liver image, the classification of lesions depends heavily on the characteristics of the lesions including internal echo, morphology, edge, echogenicity, and posterior echo enhancement. These characteristics are differently observed according to ROI selection methods that may indeed significantly impact the classification performances. Currently developed ROI selection methods have limitation for guaranteeing robust classification performance for focal liver lesions, mainly due to the inherent difficulties that represent all ultrasonic appearances of characteristics of lesion. In order to obtain better and more stable classification performances, we propose a new and novel approach, so-called multiple-ROI based focal liver lesion classification. The proposed approach properly combines the advantages of existing ROI selection methods to represent well various ultrasonic appearances of liver lesions including internal echo, morphology, edge, echogenicity, and posterior echo enhancement. To verify the effectiveness of the proposed ROI selection approach, extensive and comparative experiments have been performed using a total of 150 ultrasound images. Each ultrasound image contains one corresponding focal liver lesion so that a total of 150 focal liver lesions is used, comprising of 50 cysts, 50 hemangiomas, and 50 malignancies. Experimental results show that the proposed multiple-ROI-based approach can achieve the enhanced and stable classification performance regardless of features being used. In addition, our proposed method outperforms other existing classification methods designed for focal liver lesion classification. Especially, the proposed approach attains classification accuracy of up to 80% over well-known challenging task of classifying the hemangiomas and malignancies.     

基于局部灰度最大和改进Mahalanobis距离分类的肺结节检测算法

CT图像中肺结节检测一直是肺癌CAD系统的关键和难点。提出了一种孤立性肺结节自动检测算法,首先对原始CT图像进行有效、准确的肺实质分割;采用寻找局部灰度最大值方法对ROI进行初始分割;再对分割出的各ROI进行特征提取,利用SVM方法对每个特征进行定量描述,根据SVM单特征分类准确率对Mahalanobis距离进行加权改进,最后采用基于改进的Mahalanobis距离进行肺结节分类。实验结果表明,该算法可以较好地提取出CT图像中的孤立性肺结节,具有较高的灵敏度和较低的漏诊率,可以为医生诊断早期肺癌病灶提供帮助信息。