Symtosis: A liver ultrasound tissue characterization and risk stratification in optimized deep learning paradigm
Symtosis: A liver ultrasound tissue characterization and risk stratification in optimized deep learning paradigm
复制标题
DOI：
10.1016/j.cmpb.2017.12.016
复制
发表时间：
2018-03-01
影响因子：
6.1
通讯作者：
Suri, Jasjit S.
中科院分区：
文献类型：
--
作者：
Biswas, Mainak;Kuppili, Venkatanareshbabu;Suri, Jasjit S.
研究背景与目的脂肪性肝病（Fatty Liver Disease，FLD）是由肝细胞内脂肪沉积引起的疾病，是肝癌等晚期疾病的前身。应用于FLD检测和使用超声（US）的风险分层的机器学习（ML）技术在计算组织表征特征方面具有局限性，从而限制了准确性。这项研究提出了一种基于深度学习（DL）的范式，当在交叉过程中通过22层神经网络时，该范式可以计算每张图像近700万个权重。验证（培训和测试）范式。DL架构由级联的操作层组成，例如：卷积，池化，整流线性单元，dropout和一个称为inception model的特殊块，它提供了速度和效率。所有数据分析都是在通过去除背景信息获得的优化组织区域中进行的。我们基准的DL系统对传统的ML协议：支持向量机（SVM）和极端学习机（ELM）。ResultsThe肝脏US数据包括63例（27正常/36异常）。使用K10交叉验证协议（90%训练和10%测试），SVM、ELM和DL系统的检测和风险分层准确率分别为82%、92%和100%。相应的曲线下面积分别为：0.79、0.92和1.0。结论与传统的机器学习系统SVM和ELM相比，DL系统在肝脏检测和风险分层方面具有上级性能。
Background and ObjectiveFatty Liver Disease (FLD) - a disease caused by deposition of fat in liver cells, is predecessor to terminal diseases such as liver cancer. The machine learning (ML) techniques applied for FLD detection and risk stratification using ultrasound (US) have limitations in computing tissue characterization features, thereby limiting the accuracy.MethodsUnder the class of Symtosis for FLD detection and risk stratification, this study presents a Deep Learning (DL)-based paradigm that computes nearly seven million weights per image when passed through a 22 layered neural network during the cross-validation (training and testing) paradigm. The DL architecture consists of cascaded layers of operations such as: convolution, pooling, rectified linear unit, dropout and a special block called inception model that provides speed and efficiency. All data analysis is performed in optimized tissue region, obtained by removing background information. We benchmark the DL system against the conventional ML protocols: support vector machine (SVM) and extreme learning machine (ELM).ResultsThe liver US data consists of 63 patients (27 normal/36 abnormal). Using the K10 cross-validation protocol (90% training and 10% testing), the detection and risk stratification accuracies are: 82%, 92% and 100% for SVM, ELM and DL systems, respectively. The corresponding area under the curve is: 0.79, 0.92 and 1.0, respectively. We further validate our DL system using two class biometric facial data that yields an accuracy of 99%.ConclusionDL system shows a superior performance for liver detection and risk stratification compared to conventional machine learning systems: SVM and ELM.