Faculty Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK.
Faculty of Engineering, University of Glasgow, Glasgow, UK.
Abdom Radiol (NY). 2024 Feb;49(2):651-661. doi: 10.1007/s00261-023-04147-x. Epub 2024 Jan 12.
Abdominal ultrasound screening requires the capture of multiple standardized plane views as per clinical guidelines. Currently, the extent of adherence to such guidelines is dependent entirely on the skills of the sonographer. The use of neural network classification has the potential to better standardize captured plane views and streamline plane capture reducing the time burden on operators by combatting operator variability.
A dataset consisting of 16 routine upper abdominal ultrasound scans from 64 patients was used to test the classification accuracy of 9 neural networks. These networks were tested on both a small, idealised subset of 800 samples as well as full video sweeps of the region of interest using stratified sampling and transfer learning.
The highest validation accuracy attained by both GoogLeNet and InceptionV3 is 83.9% using transfer learning and the large sample set of 26,294 images. A top-2 accuracy of 95.1% was achieved using InceptionV3. Alexnet attained the highest accuracy of 79.5% (top-2 of 91.5%) for the smaller sample set of 800 images. The neural networks evaluated during this study were also successfully able to identify problematic individual cross sections such as between kidneys, with right and left kidney being accurately identified 78.6% and 89.7%, respectively.
Dataset size proved a more important factor in determining accuracy than network selection with more complex neural networks providing higher accuracy as dataset size increases and simpler linear neural networks providing better results where the dataset is small.
腹部超声筛查需要根据临床指南获取多个标准化的平面视图。目前，对这些指南的遵守程度完全取决于超声医师的技能。神经网络分类的使用有可能更好地规范所捕获的平面视图，并简化平面捕获，通过对抗操作人员的可变性来减少操作人员的时间负担。
使用包含 64 名患者的 16 例常规上腹部超声扫描的数据集来测试 9 个神经网络的分类准确性。这些网络在 800 个样本的小理想子集以及使用分层抽样和迁移学习的感兴趣区域的全视频扫描上进行了测试。
使用迁移学习和 26294 张图像的大样本集，GoogLeNet 和 InceptionV3 获得的最高验证准确率均为 83.9%。使用 InceptionV3 实现了 95.1%的最高准确率。对于 800 个图像的较小样本集，Alexnet 达到了最高的准确率 79.5%（前 2 名中的 91.5%）。在这项研究中评估的神经网络还成功地能够识别出有问题的个别横截段，例如肾脏之间，左右肾脏的识别准确率分别为 78.6%和 89.7%。
数据集大小是确定准确性的更重要因素，而不是网络选择，随着数据集大小的增加，更复杂的神经网络提供更高的准确性，而简单的线性神经网络在数据集较小时提供更好的结果。
