Pattern Recognition Lab, Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany.
Department of Oral and Maxillofacial Surgery, University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany.
Int J Comput Assist Radiol Surg. 2019 Jan;14(1):31-42. doi: 10.1007/s11548-018-1836-1. Epub 2018 Aug 4.
Probe-based confocal laser endomicroscopy (pCLE) is a subcellular in vivo imaging technique capable of producing images that enable diagnosis of malign structural modifications in epithelial tissue. Images acquired with pCLE are, however, often tainted by significant artifacts that impair diagnosis. This is especially detrimental for automated image analysis, which is why said images are often excluded from recognition pipelines.
We present an approach for the automatic detection of motion artifacts in pCLE images and apply this methodology to a data set of 15 thousand images of epithelial tissue acquired in the oral cavity and the vocal folds. The approach is based on transfer learning from intermediate endpoints within a pre-trained Inception v3 network with tailored preprocessing. For detection within the non-rectangular pCLE images, we perform pooling within the activation maps of the network and evaluate this at different network depths.
We achieved area under the ROC curve values of 0.92 with the proposed method, compared to 0.80 for the best feature-based machine learning approach. Our overall accuracy with the presented approach is 94.8%.
Over traditional machine learning approaches with state-of-the-art features, we achieved significantly improved overall performance.
基于探针的共聚焦激光显微内镜（pCLE）是一种能够产生图像的亚细胞体内成像技术，这些图像可用于诊断上皮组织中的恶性结构改变。然而，pCLE 采集的图像常常受到严重伪影的影响，从而影响诊断。这对于自动化图像分析尤其不利，这就是为什么这些图像通常被排除在识别管道之外的原因。
我们提出了一种用于自动检测 pCLE 图像中运动伪影的方法，并将该方法应用于在口腔和声带采集的 15000 张上皮组织图像的数据集。该方法基于从经过预训练的 Inception v3 网络的中间端点进行迁移学习，并进行了针对性的预处理。对于非矩形 pCLE 图像中的检测，我们在网络的激活图中执行池化，并在不同的网络深度进行评估。
与基于最佳特征的机器学习方法的 0.80 相比，我们提出的方法在 ROC 曲线下的面积达到了 0.92。我们提出的方法的总体准确率为 94.8%。
与具有最新特征的传统机器学习方法相比，我们实现了显著提高的整体性能。
