Suppr超能文献

介入放射学的语音与运动控制:要求与可行性

Speech and motion control for interventional radiology: requirements and feasibility.

作者信息

Hötker Andreas M, Pitton Michael B, Mildenberger Peter, Düber Christoph

机构信息

Department of Diagnostic and Interventional Radiology, Universitätsmedizin Mainz, Langenbeckstr. 1, 55131 , Mainz, Germany,

出版信息

Int J Comput Assist Radiol Surg. 2013 Nov;8(6):997-1002. doi: 10.1007/s11548-013-0841-7. Epub 2013 Apr 13.

Abstract

PURPOSE

Interventional radiology is performed in a sterile environment where speech and motion control of image review is needed to simplify and expedite routine procedures. The requirements and limitations were defined by testing an interventional radiology test bed speech and motion control system.

METHODS

Motion control software was implemented using the Microsoft[Formula: see text] Kinect[Formula: see text] (Microsoft Corp., USA) framework. The system was tested by 10 participants using a predefined set of six voice and six gesture commands under different lighting conditions to assess the influence of illumination on command recognition. The participants rated the convenience of the application and its possible use in everyday clinical routine. A basic set of voice or gesture commands required for interventional radiology were identified.

RESULTS

The majority (93 %) of commands were recognized successfully. Speech commands were less prone to errors than gesture commands. Unwanted side effects occurred (e.g., accidentally issuing a gesture command) in about 30 % of cases. Dimmed lighting conditions did not have a measurable effect on the recognition rate. Six out of 10 participants would consider using the application in everyday routine. The necessary voice/gesture commands for interventional radiology were identified and integrated into the control system.

CONCLUSION

Speech and motion control of image review provides a new man-machine interface for radiological image handling that is especially useful in sterile environments due to no-touch navigation. Command recognition rates were high and remained stable under different lighting conditions. However, the rate of accidental triggering due to unintended commands should be reduced.

摘要

目的

介入放射学操作在无菌环境中进行,在此环境下需要对图像查看进行语音和动作控制,以简化和加快常规操作流程。通过测试介入放射学试验台语音和动作控制系统来明确其要求和局限性。

方法

使用微软公司的Kinect框架实现动作控制软件。10名参与者在不同光照条件下,使用预定义的6个语音命令和6个手势命令对该系统进行测试,以评估光照对命令识别的影响。参与者对该应用的便利性及其在日常临床操作中的可能用途进行评分。确定了介入放射学所需的一组基本语音或手势命令。

结果

大部分(93%)命令被成功识别。语音命令比手势命令更不易出错。约30%的情况下会出现意外副作用(如意外发出手势命令)。光线昏暗的条件对识别率没有可测量的影响。10名参与者中有6人会考虑在日常操作中使用该应用。确定了介入放射学所需的语音/手势命令并将其集成到控制系统中。

结论

图像查看的语音和动作控制为放射图像处理提供了一种新的人机界面,由于无需接触式导航,在无菌环境中特别有用。命令识别率很高,并且在不同光照条件下保持稳定。然而,应降低因意外命令导致的意外触发率。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验