目的:目的是将先前验证的深度学习算法应用于新的甲状腺结节超声图像数据集,并将其性能与放射科医生进行比较。方法:先前的研究提出了一种能够检测甲状腺结节,然后使用两个超声图像进行恶性分类的算法。从1278个结节训练了多任务深度卷积神经网络,最初用99个单独的结节进行了测试。结果与放射科医生相当。与培训案例相比,使用来自不同制造商和产品类型的超声计算机成像的378个结节进一步测试了该算法。要求四名经验丰富的放射科医生评估结节,以与深度学习进行比较。结果:用参数,二维估计计算了深度学习算法和四个放射科医生的曲线(AUC)面积。对于深度学习算法,AUC为0.70(95%CI:0.64-0.75)。放射科医生的AUC为0.66(95%CI:0.61-0.71),0.67(95%CI:0.62-0.73),0.68(95%CI:0.63-0.73)和0.66(95%CI:95%CI:0.61-0.71)。结论:在新的测试数据集中,深度学习算法与所有四个放射科医生都达到了类似的性能。
translated by 谷歌翻译
使用Kellgren-Lawence分级系统在放射线照片中评估放射性骨关节炎的严重程度评估放射科医生的表现,是放射学家的表现。根据Kellgren-Lawence分级系统,开发一种自动化的基于深度学习的算法,该算法使用膝盖X光片的后侧(PA)和侧面(LAT)视图来评估膝关节骨关节炎的严重程度。我们使用了来自多中心骨关节炎研究的2802名患者的9739例检查的数据集(大多数)。该数据集分为2040名患者的训练集,259例患者的验证和503例患者的测试组。一种新型的基于深度学习的方法用于评估膝关节OA分为两个步骤:(1)图像中膝关节的定位,(2)根据KL分级系统进行分类。我们的方法同时使用PA和LAT视图作为模型的输入。将算法生成的分数与整个测试集的最多数据集中提供的等级以及我们机构中5位放射科医生提供的成绩进行了比较。与大多数数据集中提供的评分相比,该模型在整个测试集上获得了71.90%的多级准确性。该组的二次加权KAPPA系数为0.9066。我们机构的所有放射科医生对研究的平均二次加权Kappa为0.748。我们机构的算法和放射科医生之间的平均二次加权Kappa为0.769。所提出的模型表明,KL分类与MSK放射科医生的等效性,但显然可重复性。我们的模型还与我们机构的放射科医生同意与放射科医生相同的程度。该算法可用于提供膝关节炎严重程度的可重复评估。
translated by 谷歌翻译
歧管假设是深度学习成功背后的核心机制,因此了解图像数据的内在流形结构对于研究神经网络如何从数据中学习至关重要。固有的数据集歧管及其与学习难度的关系最近开始研究自然图像的共同领域,但是几乎没有尝试进行放射学图像的研究。我们在这里解决这个问题。首先,我们比较放射学和自然图像的固有歧管维度。我们还研究了固有维度和泛化能力之间的关系。我们的分析表明,自然图像数据集通常比放射学图像具有更高数量的固有维度。但是,对医学图像的概括能力与内在维度之间的关系更加牢固,这可以解释为具有固有特征的放射学图像更难学习。这些结果为直觉提供了更具原则性的基础,即放射学图像要比机器学习研究所共有的自然图像数据集更具挑战性。我们认为,与其直接将为自然图像开发的模型直接应用于放射成像领域,不应对开发更适合该域的特定特征定制的体系结构和算法进行更多的注意。我们的论文中显示的研究表明了这些特征以及与自然图像的差异,是该方向上的重要第一步。
translated by 谷歌翻译
膝关节X射线上的膝盖骨关节炎(KOA)的评估是使用总膝关节置换术的中心标准。但是,该评估遭受了不精确的标准,并且读取器间的可变性非常高。对KOA严重性的算法,自动评估可以通过提高其使用的适当性来改善膝盖替代程序的总体结果。我们提出了一种基于深度学习的新型五步算法,以自动从X光片后验(PA)视图对KOA进行评级:(1)图像预处理(2)使用Yolo V3-tiny模型,图像在图像中定位膝关节, (3)使用基于卷积神经网络的分类器对骨关节炎的严重程度进行初步评估,(4)关节分割和关节空间狭窄(JSN)的计算(JSN)和(5),JSN和最初的结合评估确定最终的凯尔格伦法律(KL)得分。此外,通过显示用于进行评估的分割面具,我们的算法与典型的“黑匣子”深度学习分类器相比表现出更高的透明度。我们使用我们机构的两个公共数据集和一个数据集进行了全面的评估,并表明我们的算法达到了最先进的性能。此外,我们还从机构中的多个放射科医生那里收集了评分,并表明我们的算法在放射科医生级别进行。该软件已在https://github.com/maciejmazurowowski/osteoarthitis-classification上公开提供。
translated by 谷歌翻译
我们对标准物体检测模型的特征金字塔网络进行了改进。我们通过本地图像翻译和密切地调用我们的方法增强功能金字塔网络。副本通过同时(1)产生逼真但具有模拟对象的假图像来提高对象检测性能,以减轻注意力机制的数据饥饿问题,并通过新颖的对图像特征贴片进行注意力推进检测模型架构。具体地,我们使用卷积AutomEncoder作为生成器来通过本地插值将对象注入图像来创建新图像,并在隐藏层中提取的功能重建。然后由于模拟图像数量较多,我们使用可视变压器来增强每个Reset层的输出,该层用作特征金字塔网络的输入。我们将方法应用于检测数字乳房断层合成扫描(DBT)中的病变,高分辨率医学成像模塑在乳腺癌筛查中至关重要。我们在定性和定量上展示复制品可以通过实验结果利用增强的标准物体检测框架提高肿瘤检测的准确性。
translated by 谷歌翻译
目的:为全身CT设计多疾病分类扫描使用自动提取标签从放射科文reports.Materials和方法三个不同的器官系统:这项回顾性研究共有12,092例患者(平均年龄57 + - 18; 6172名妇女)包括对模型开发和测试(2012-2017自)。基于规则的算法被用来从12,092患者提取13667身体CT扫描19,225疾病的标签。使用三维DenseVNet,三个器官系统是分段的:肺和胸膜;肝胆;和肾脏及输尿管。对于每个器官,三维卷积神经网络分类没有明显的疾病与四种常见疾病为跨越所有三个模型总共15个不同的标签。测试是在相对于2875个手动导出的参考标签2158个CT体积的子集从2133名患者( - ; 1079名妇女18,平均年龄58 +)进行。性能报告为曲线(AUC)与通过方法德朗95%置信区间下接收器的操作特性的区域。结果:提取的标签说明书验证确认91%横跨15个不同的唱片公司99%的准确率。对于肺和胸膜标签的AUC分别为:肺不张0.77(95%CI:0.74,0.81),结节0.65(0.61,0.69),肺气肿0.89(0.86,0.92),积液0.97(0.96,0.98),并且没有明显的疾病0.89( 0.87,0.91)。对于肝和胆囊的AUC分别为:肝胆钙化0.62(95%CI:0.56,0.67),病变0.73(0.69,0.77),扩张0.87(0.84,0.90),脂肪0.89(0.86,0.92),并且没有明显的疾病0.82( 0.78,0.85)。对于肾脏及输尿管的AUC分别为:石0.83(95%CI:0.79,0.87),萎缩0.92(0.89,0.94),病变0.68(0.64,0.72),囊肿0.70(0.66,0.73),并且没有明显的疾病0.79(0.75 ,0.83)。结论:弱监督深度学习模型能够在多器官系统不同的疾病分类。
translated by 谷歌翻译
In this study, we systematically investigate the impact of class imbalance on classification performance of convolutional neural networks (CNNs) and compare frequently used methods to address the issue. Class imbalance is a common problem that has been comprehensively studied in classical machine learning, yet very limited systematic research is available in the context of deep learning. In our study, we use three benchmark datasets of increasing complexity, MNIST, CIFAR-10 and ImageNet, to investigate the effects of imbalance on classification and perform an extensive comparison of several methods to address the issue: oversampling, undersampling, two-phase training, and thresholding that compensates for prior class probabilities. Our main evaluation metric is area under the receiver operating characteristic curve (ROC AUC) adjusted to multi-class tasks since overall accuracy metric is associated with notable difficulties in the context of imbalanced data. Based on results from our experiments we conclude that (i) the effect of class imbalance on classification performance is detrimental; (ii) the method of addressing class imbalance that emerged as dominant in almost all analyzed scenarios was oversampling; (iii) oversampling should be applied to the level that completely eliminates the imbalance, whereas the optimal undersampling ratio depends on the extent of imbalance; (iv) as opposed to some classical machine learning models, oversampling does not cause overfitting of CNNs; (v) thresholding should be applied to compensate for prior class probabilities when overall number of properly classified cases is of interest.
translated by 谷歌翻译
This short report reviews the current state of the research and methodology on theoretical and practical aspects of Artificial Neural Networks (ANN). It was prepared to gather state-of-the-art knowledge needed to construct complex, hypercomplex and fuzzy neural networks. The report reflects the individual interests of the authors and, by now means, cannot be treated as a comprehensive review of the ANN discipline. Considering the fast development of this field, it is currently impossible to do a detailed review of a considerable number of pages. The report is an outcome of the Project 'The Strategic Research Partnership for the mathematical aspects of complex, hypercomplex and fuzzy neural networks' meeting at the University of Warmia and Mazury in Olsztyn, Poland, organized in September 2022.
translated by 谷歌翻译
Transfer learning is a popular technique for improving the performance of neural networks. However, existing methods are limited to transferring parameters between networks with same architectures. We present a method for transferring parameters between neural networks with different architectures. Our method, called DPIAT, uses dynamic programming to match blocks and layers between architectures and transfer parameters efficiently. Compared to existing parameter prediction and random initialization methods, it significantly improves training efficiency and validation accuracy. In experiments on ImageNet, our method improved validation accuracy by an average of 1.6 times after 50 epochs of training. DPIAT allows both researchers and neural architecture search systems to modify trained networks and reuse knowledge, avoiding the need for retraining from scratch. We also introduce a network architecture similarity measure, enabling users to choose the best source network without any training.
translated by 谷歌翻译
Insufficient image spatial resolution is a serious limitation in many practical scenarios, especially when acquiring images at a finer scale is infeasible or brings higher costs. This is inherent to remote sensing, including Sentinel-2 satellite images that are available free of charge at a high revisit frequency, but whose spatial resolution is limited to 10 m ground sampling distance. The resolution can be increased with super-resolution algorithms, in particular when performed from multiple images captured at subsequent revisits of a satellite, taking advantage of information fusion that leads to enhanced reconstruction accuracy. One of the obstacles in multi-image super-resolution consists in the scarcity of real-world benchmarks - commonly, simulated data are exploited which do not fully reflect the operating conditions. In this paper, we introduce a new MuS2 benchmark for super-resolving multiple Sentinel-2 images, with WorldView-2 imagery used as the high-resolution reference. Within MuS2, we publish the first end-to-end evaluation procedure for this problem which we expect to help the researchers in advancing the state of the art in multi-image super-resolution.
translated by 谷歌翻译