arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2303.09648cs.CV
收藏

移位窗Transformers在脑动脉瘤显微外科诊断中的应用

Shifted-Windows Transformers for the Detection of Cerebral Aneurysms in Microsurgery

University of Michigan(密歇根大学)

Jinfan Zhou, William Muirhead, Simon C. Williams, Danail Stoyanov, Hani J. Marcus, Evangelos B. Mazomenos

摘要:

目的:显微外科动脉瘤夹闭手术(MACS)具有术中动脉瘤破裂的高风险。当动脉瘤在手术视频中暴露时,自动识别实例将是神经导航的有价值的参考点,指示相变,更重要的是指定破裂的高风险时刻。本文介绍了包含16个具有帧级专家注释的手术视频的MACS数据集,并提出了一种用于手术场景理解的学习方法,以识别手术显微镜视野中存在动脉瘤的视频帧。研究方法:尽管数据集不平衡(80%不存在,20%存在)并且在没有明确注释的情况下开发,但我们证明了基于transformer的深度学习架构(MACSSwin-T,vidMACSSwin-T)检测动脉瘤并相应地对MACS帧进行分类的适用性。我们评估了所提出的模型在多重交叉验证实验与独立集和一组看不见的15张图像对10个人类专家(神经外科医生)。结果如下:图像和视频级方法的平均(跨折叠)准确率分别为80.8%(范围78.5%-82.4%)和87.1%(范围85.1%-91.3%),表明模型有效地学习分类任务。对模型类激活图的定性评价显示,这些激活图定位于动脉瘤的实际位置。根据决策阈值的不同,MACSWin-T在看不见的图像中实现了66.7%到86.7%的准确率,而人类评分员的准确率为82%,具有中度到强相关性。

英文摘要:

Purpose: Microsurgical Aneurysm Clipping Surgery (MACS) carries a high risk for intraoperative aneurysm rupture. Automated recognition of instances when the aneurysm is exposed in the surgical video would be a valuable reference point for neuronavigation, indicating phase transitioning and more importantly designating moments of high risk for rupture. This article introduces the MACS dataset containing 16 surgical videos with frame-level expert annotations and proposes a learning methodology for surgical scene understanding identifying video frames with the aneurysm present in the operating microscope's field-of-view. Methods: Despite the dataset imbalance (80% no presence, 20% presence) and developed without explicit annotations, we demonstrate the applicability of Transformer-based deep learning architectures (MACSSwin-T, vidMACSSwin-T) to detect the aneurysm and classify MACS frames accordingly. We evaluate the proposed models in multiple-fold cross-validation experiments with independent sets and in an unseen set of 15 images against 10 human experts (neurosurgeons). Results: Average (across folds) accuracy of 80.8% (range 78.5%-82.4%) and 87.1% (range 85.1%-91.3%) is obtained for the image- and video-level approach respectively, demonstrating that the models effectively learn the classification task. Qualitative evaluation of the models' class activation maps show these to be localized on the aneurysm's actual location. Depending on the decision threshold, MACSWin-T achieves 66.7% to 86.7% accuracy in the unseen images, compared to 82% of human raters, with moderate to strong correlation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑