彭进业^1，2，*， 杨瑞靖^1，， 冯晓毅^2，

（  1、西北大学信息科学与技术学院；       2、西北工业大学电子信息学院；  ）

摘要： 人脸表情被认为是表达内在情感的一种标志。在过去几十年间，自动识别自发的面部表情成为了备受关注的研究热点。与此同时，在自发的面部表情中自动识别疼痛的表情也越来越引起研究者的关注，因为这将可以广泛应用于医院等医疗机构中。比如，针对那些无法用语言表达疼痛的病人，如新生儿，患有痴呆症的患者，以及被重症监护的病人等等。有许多研究工作致力于基于人的面部表情来识别疼痛的二值结果（如疼，或者不疼）。然而，这样的结果不能够充分反应病人的疼痛程度而且可能导致不充足的治疗。此外，从静态的图像中检测疼痛会丢失很多有价值的信息。因此，在视频中自动检测患者的疼痛程度这一课题在领域中越来越被学者所重视。然而，目前研究的主要问题集中在缺乏在真实场景中检测疼痛的数据集，面部特征数据的降维等方面。本文回顾了有关疼痛检测的现有数据集，现有的疼痛检测的一般模型和方法，以及目前存在的挑战。
关键词： 模式识别；人脸识别；疼痛检测

Peng Jinye^1,*， Yang Ruijing^2,， Feng Xiaoyi^1,

（  1、School of Information Science and Technology, Northwest University, 710127；       2、 School of Electronics and Information, Northwestern Polytechnical University, 710072；  ）

Abstract： Facial expression is deemed as a sign of expressing one's internal emotion. Automatic recognition of spontaneous facial expression is the most popular research area during the past decades. Similarly, automatically recognizing pain from spontaneous facial expression is of increased attention, because of its wide application in clinical settings, such as for patients who cannot express their pain vocally. Such special categories of patients may be neonates, patients with dementia, patients in intensive care units, etc. Lots of research work focus on detecting binary results of pain from facial expressions, whereas this insufficient results may lead to under treatment for patients. In addition, detecting pain from static imges may lose lots of valuable information. Therefore, automatic detection of pain intensity in video sequences has become popular in the field. However, the lack of databases in natural scenarios related to pain expression and the reduction of the large dimensionality of facial features are the major problem in this field. In the present paper, several databases and the basic detecting model and methods that are used in concurrent studies are reviewed , together with the remaining challenges of pain detection.

Tag：

点此返回栏目查看更多>>>参考论文