An intelligent pain recognition system based on facial expressions and pysiological signals

Abstract

In current medical care, the most common method of pain assessment is based on patients' self-assessments. People's pain sensation and tolerance threshold are different, making it almost impossible for clinicians to assess the degree of pain accurately and objectively in patients. An accurate assessment of pain is essential for subsequent treatment. This work provides a system for automatically identifying pain levels based on an open-source BioVid heat pain database. Skin conductance characteristics at different pain levels were analyzed. The performance is tested for models based on facial expression and electrocardiogram (ECG) data models. The results show that facial expressions and ECG can provide feedback on the patient's current pain level based on an accurate analysis of different pain levels and baselines. This can significantly help objectively assess a patient's pain level.--Author's abstract