Jiang Jiaxin, Jiang Xiaoya, Xu Lei, Zhang Yan, Zheng Yuwen, Kong Dexing
School of Mathematical Sciences, Zhejiang University, Hangzhou, China.
Department of Mathematics, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong SAR, China.
Front Oncol. 2023 Jun 22;13:1177225. doi: 10.3389/fonc.2023.1177225. eCollection 2023.
Deep learning technology has been widely applied to medical image analysis. But due to the limitations of its own imaging principle, ultrasound image has the disadvantages of low resolution and high Speckle Noise density, which not only hinder the diagnosis of patients' conditions but also affect the extraction of ultrasound image features by computer technology.
In this study, we investigate the robustness of deep convolutional neural network (CNN) for classification, segmentation, and target detection of breast ultrasound image through random Salt & Pepper Noise and Gaussian Noise.
We trained and validated 9 CNN architectures in 8617 breast ultrasound images, but tested the models with noisy test set. Then, we trained and validated 9 CNN architectures with different levels of noise in these breast ultrasound images, and tested the models with noisy test set. Diseases of each breast ultrasound image in our dataset were annotated and voted by three sonographers based on their malignancy suspiciousness. we use evaluation indexes to evaluate the robustness of the neural network algorithm respectively.
There is a moderate to high impact (The accuracy of the model decreased by about 5%-40%) on model accuracy when Salt and Pepper Noise, Speckle Noise, or Gaussian Noise is introduced to the images respectively. Consequently, DenseNet, UNet++ and Yolov5 were selected as the most robust model based on the selected index. When any two of these three kinds of noise are introduced into the image at the same time, the accuracy of the model will be greatly affected.
Our experimental results reveal new insights: The variation trend of accuracy with the noise level in Each network used for classification tasks and object detection tasks has some unique characteristics. This finding provides us with a method to reveal the black-box architecture of computer-aided diagnosis (CAD) systems. On the other hand, the purpose of this study is to explore the impact of adding noise directly to the image on the performance of neural networks, which is different from the existing articles on robustness in the field of medical image processing. Consequently, it provides a new way to evaluate the robustness of CAD systems in the future.
深度学习技术已广泛应用于医学图像分析。但由于超声图像自身成像原理的限制，存在分辨率低、斑点噪声密度高的缺点，这不仅阻碍了对患者病情的诊断，也影响了计算机技术对超声图像特征的提取。
在本研究中，我们通过随机椒盐噪声和高斯噪声来研究深度卷积神经网络（CNN）对乳腺超声图像进行分类、分割和目标检测的鲁棒性。
我们在8617幅乳腺超声图像中训练并验证了9种CNN架构，但使用有噪声的测试集对模型进行测试。然后，我们在这些乳腺超声图像中使用不同噪声水平训练并验证9种CNN架构，并使用有噪声的测试集对模型进行测试。我们数据集中每幅乳腺超声图像的疾病由三位超声医师根据其恶性可疑程度进行标注和投票。我们分别使用评估指标来评估神经网络算法的鲁棒性。
当分别向图像中引入椒盐噪声、斑点噪声或高斯噪声时，对模型准确率有中度到高度的影响（模型准确率下降约5%-40%）。因此，基于所选指标，DenseNet、UNet++和Yolov5被选为最鲁棒的模型。当同时将这三种噪声中的任意两种引入图像时，模型的准确率将受到极大影响。
我们的实验结果揭示了新的见解：用于分类任务和目标检测任务的每个网络中，准确率随噪声水平的变化趋势具有一些独特的特征。这一发现为我们提供了一种揭示计算机辅助诊断（CAD）系统黑箱架构的方法。另一方面，本研究的目的是探索直接向图像中添加噪声对神经网络性能的影响，这与医学图像处理领域现有关于鲁棒性的文章不同。因此，它为未来评估CAD系统的鲁棒性提供了一种新方法。
