Pain level evaluation method based on facial pain expression video

Abstract

The invention discloses a pain level evaluation method based on a facial pain expression video. The method is characterized by comprising a training stage and a recognition stage, wherein the trainingstage comprises the following steps: (1) extracting a human face pain expression video from a human face pain expression video library, and editing according to a set time period and different pain levels to obtain a plurality of levels of video clips with different pain levels; (2) preprocessing the video clips with different pain levels, and training and constructing a pain level evaluation network comprising a convolutional neural network and a long-short-term memory network; the recognition stage comprises the following steps: collecting a human face pain expression video clip of a patient in a set time period, taking the human face pain expression video clip as input of a pain level evaluation network, and evaluating and outputting a pain level value corresponding to the human face pain expression video clip by the pain level evaluation network. The pain level evaluation method based on the facial pain expression video is high in accuracy and robustness.

Description

technical field [0001] The invention relates to the technical fields of image recognition and deep learning, in particular to a pain level assessment method based on a human face pain expression video. Background technique [0002] The World Health Organization clearly stated in 2000 that pain is a kind of human disease, often manifested as an unpleasant emotion, and accompanied by potential tissue damage. In 2001, the World Health Organization (WHO) officially recognized pain as the "fifth vital sign" after the four vital signs of breathing, pulse, body temperature and blood pressure, and pointed out that pain relief is a basic human right. Assessing pain scientifically is the first step for clinicians to implement standardized treatment of pain. Usually, the commonly used clinical assessment of pain level is Visual Analogue Scale (VAS), Numerical Pain Scale (NRS), Wong-Baker Face Scale (FRS) and so on. One of the main bases of these scoring methods is the continuous obse...

AI Technical Summary

Problems solved by technology

Affected by light, head posture, image background, etc., the evaluation of the patient's pain level based on a static picture is easy to fluctuate, and it is difficult to accurately reflect the patient's pain level

Images