Haidegger Tamás, Nagy Melinda, Lehotsky Akos, Szilágyi László
Budapest University of Technology and Economics, Dept. of Control Engineering and Information Technology, Budapest, Hungary.
Med Image Comput Comput Assist Interv. 2011;14(Pt 3):619-26. doi: 10.1007/978-3-642-23626-6_76.
Nosocomial infections are the undesirable result of a treatment in a hospital, or a health care service unit, not related to the patient's original condition. Despite the evolution of medicine, fundamental problems with hand hygiene remain existent, leading to the spread of nosocomial infections. Our group has been working on a generic solution to provide a method and apparatus to teach and verify proper hand disinfection. The general idea is to mark the skin surfaces that were sufficiently treated with alcoholic hand rub. Digital image processing is employed to determine the location of these areas and overlay it on the segmented hand, visualizing the results in an intuitive form. A non-disruptive ultraviolet marker is mixed to a commercially available hand rub, therefore leaving the original hand washing workflow intact. Digital images are taken in an enclosed device we developed for this purpose. First, robust hand contour segmentation is performed, then a histogram-based formulation of the fuzzy c-means algorithm is employed for the classification of clean versus dirty regions, minimizing the processing time of the images. The method and device have been tested in 3 hospitals in Hungary, Romania and Singapore, on surgeons, residents, medical students and nurses. A health care professional verified the results of the segmentation, since no gold standard is available for the recorded human cases. We were able to identify the hand boundaries correctly in 99.2% of the cases. The device can give objective feedback to medical students and staff to develop and maintain proper hand disinfection practice.
医院感染是在医院或医疗服务单位进行治疗时出现的不良结果，与患者的原始病情无关。尽管医学不断发展，但手部卫生的基本问题仍然存在，导致医院感染的传播。我们的团队一直在致力于提供一种通用解决方案，以提供一种教导和验证正确手部消毒的方法和设备。总体思路是标记用酒精擦手液充分处理过的皮肤表面。利用数字图像处理来确定这些区域的位置，并将其叠加在分割后的手上，以直观的形式呈现结果。一种无干扰的紫外线标记物被混入市售的擦手液中，因此保持了原有的洗手流程不变。数字图像是在我们为此专门开发的封闭设备中拍摄的。首先，进行稳健的手部轮廓分割，然后采用基于直方图的模糊c均值算法对清洁区域和脏污区域进行分类，以尽量减少图像的处理时间。该方法和设备已在匈牙利、罗马尼亚和新加坡的3家医院对外科医生、住院医生、医学生和护士进行了测试。由于没有针对所记录的人类病例的金标准，因此由一名医疗保健专业人员对分割结果进行了验证。在99.2%的病例中，我们能够正确识别手部边界。该设备可以向医学生和工作人员提供客观反馈，以培养和保持正确的手部消毒习惯。
