基于预训练 DenseNet201 与胶囊网络的肩部植入物制造商分类。

Classification of Shoulder Implant Manufacturer Using Pre-Trained DenseNet201 Combined With Capsule Network.

机构信息

School of Optoelectronic Information and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, China.

Reproductive Medicine Center, Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine, Shanghai, China.

出版信息

Int J Med Robot. 2024 Oct;20(5):e2672. doi: 10.1002/rcs.2672.

DOI:10.1002/rcs.2672

PMID:39334563

Abstract

BACKGROUND

This study aims to accelerate revision surgery and treatment using X-ray imaging and deep learning to identify shoulder implant manufacturers in advance.

METHODS

A feature engineering approach based on principal component analysis and a k-means algorithm was used to cluster shoulder implant data. In addition, a pre-trained DenseNet201 combined with a capsule network (DenseNet201-Caps) shoulder implant classification model was proposed.

RESULTS

DenseNet201-Caps was the most effective classification model on the clustered dataset with an accuracy of 94.25% and an F1 score of 96.30%. Notably, clustering the dataset in advance improved the accuracy and the Caps implementations successfully enhanced the performance of all convolutional neural network models. The analysed results indicate that DenseNet201-Caps struggled to distinguish between the Cofield and Depuy manufacturers. Hence, a multistage classification approach was developed with an improved accuracy of 96.55% achieved.

CONCLUSIONS

The DenseNet201-Caps method enables the accurate identification of shoulder implant manufacturers.

摘要

背景

本研究旨在通过 X 射线成像和深度学习技术，提前识别肩部植入物制造商，从而加速返修手术和治疗。

方法

采用基于主成分分析和 K 均值算法的特征工程方法对肩部植入物数据进行聚类。此外，还提出了一种预训练的 DenseNet201 与胶囊网络（DenseNet201-Caps）相结合的肩部植入物分类模型。

结果

DenseNet201-Caps 在聚类数据集上的分类效果最佳，准确率为 94.25%，F1 得分为 96.30%。值得注意的是，提前对数据集进行聚类提高了模型的准确性，Caps 结构的应用成功提高了所有卷积神经网络模型的性能。分析结果表明，DenseNet201-Caps 难以区分 Cofield 和 Depuy 制造商的植入物。因此，开发了一种多阶段分类方法，准确率提高到 96.55%。