What question did this study set out to answer?

This study aims to develop a real-time pipeline for estimating Numerical Rating Scale (NRS) scores based on facial expression analysis.

May 2, 2026CureusOpen Access

Real-Time Estimation of Numerical Rating Scale (NRS) Scores Using Machine Learning-Based Facial Expression Analysis: A Proof-of-Concept Study

View Full Paper

Ask AI

Bookmark

Authors

Proof-of-concept study develops a real-time pipeline for pain assessment in patients with communication limitations, indicating feasibility.

This study aims to develop a real-time pipeline for estimating Numerical Rating Scale (NRS) scores based on facial expression analysis.
Implemented a real-time analysis pipeline using Python, integrating MediaPipe and DeepFace for facial and emotion detection.
Used seven emotion probability scores from 30 fps video streams as predictors for NRS estimation via regression models, including Random Forest.
Synthetic datasets were evaluated using leave-one-out cross-validation (LOOCV) to assess model performance.
In the pediatric dataset, the RF model achieved a Spearman’s rank correlation coefficient (ρ) of 0.7383 (p < 0.001) with a mean absolute error (MAE) of 1.5195, outpacing the baseline model (ρ = 0.2765).
In the elderly dataset, the RF model achieved ρ = 0.7566 (p < 0.001) and MAE = 1.5760.
Feature importance analysis found that 'Fear' was significant in both datasets, while 'Neutral' was notably important in the elderly dataset.

Uejima et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

May 2, 2026CureusOpen Access

View Full Paper

Ask AI

Bookmark

Authors

Proof-of-concept study develops a real-time pipeline for pain assessment in patients with communication limitations, indicating feasibility.

This study aims to develop a real-time pipeline for estimating Numerical Rating Scale (NRS) scores based on facial expression analysis.
Implemented a real-time analysis pipeline using Python, integrating MediaPipe and DeepFace for facial and emotion detection.
Used seven emotion probability scores from 30 fps video streams as predictors for NRS estimation via regression models, including Random Forest.
Synthetic datasets were evaluated using leave-one-out cross-validation (LOOCV) to assess model performance.
In the pediatric dataset, the RF model achieved a Spearman’s rank correlation coefficient (ρ) of 0.7383 (p < 0.001) with a mean absolute error (MAE) of 1.5195, outpacing the baseline model (ρ = 0.2765).
In the elderly dataset, the RF model achieved ρ = 0.7566 (p < 0.001) and MAE = 1.5760.
Feature importance analysis found that 'Fear' was significant in both datasets, while 'Neutral' was notably important in the elderly dataset.

Uejima et al. (2026) studied this question.

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark