Most of the existing pain estimation techniques depend on the response of the subject through verbal or nonverbal communication which does not suit infants, senseless and injured persons, and subjects with cognitive impairment. To bridge this gap, researchers have explored the potential of facial video- and image-based pain recognition methods. However, it provides limited classification performance with complex computation and costly frameworks including a large storage capacity. The dependence of autonomic nervous system (ANS) activities on stimulus like pain provides an alternative pathway to assess pain subjected to external stimuli through ANS-related biosignals. In this article, processing and analysis technique of electromyogram and galvanic skin response signals for assessment of pain for noncooperative subjects are presented and validated against BioVid heat pain database. Different intensities of pain are considered and characterized with the statistical features extracted from the said biosignals. It is noticed that the accuracy level of pain estimation increases with the rise in pain intensity. For highest pain level, 80% detection accuracy is achieved which outperforms the performances of facial expression-based pain assessment techniques.

中文翻译：

---

使用 EMG 和 GSR 进行疼痛评估的替代方法

大多数现有的疼痛估计技术依赖于受试者通过语言或非语言交流的反应，不适合婴儿、无知觉和受伤的人以及有认知障碍的受试者。为了弥合这一差距，研究人员探索了基于面部视频和图像的疼痛识别方法的潜力。然而，它提供了有限的分类性能，复杂的计算和昂贵的框架，包括大的存储容量。自主神经系统 (ANS) 活动对刺激（如疼痛）的依赖性提供了另一种途径来评估通过 ANS 相关生物信号受到外部刺激的疼痛。在本文中，介绍了肌电图和皮肤电反应信号的处理和分析技术，用于评估非合作受试者的疼痛，并针对 BioVid 热痛数据库进行了验证。使用从所述生物信号中提取的统计特征来考虑和表征不同强度的疼痛。值得注意的是，疼痛估计的准确度水平随着疼痛强度的增加而增加。对于最高疼痛水平，检测准确率达到 80%，优于基于面部表情的疼痛评估技术的性能。