Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 May 12.
doi: 10.1007/s10278-025-01534-2. Online ahead of print.

Deep Learning-Based Instance Appraisable Model (EDi Pain) for Pain Estimation via Facial Videos: A Retrospective Analysis and a Prospective Emergency Department Study

Yi-Cheng Yang  1 , Wen-Hsiang Cheng  1 , En-Ting Lin  1 , An-Sheng Liu  1 , Chia-Hsin Ko  2 , Chien-Hua Huang  2 , Chu-Lin Tsai #  3 , Li-Chen Fu #  4
Affiliations

Deep Learning-Based Instance Appraisable Model (EDi Pain) for Pain Estimation via Facial Videos: A Retrospective Analysis and a Prospective Emergency Department Study

Yi-Cheng Yang et al. J Imaging Inform Med. .

Abstract

Pain assessment is a critical aspect of medical care, yet automated systems for clinical pain estimation remain rare. Tools such as the visual analog scale (VAS) are commonly used in emergency departments (EDs) but rely on subjective self-reporting, with pain intensity often fluctuating during triage. An effective automated system should utilize objective labels from healthcare professionals and identify key frames from video sequences for accurate inference. In this study, short video clips were treated as instance segments for the model, with ground truth (physician-rated VAS) provided at the video level. To address the weak label problem, we proposed flexible multiple instance learning approaches. Using a specialized loss function and sampling strategy, our instance-appraisable model, EDi Pain, was trained to estimate pain intensity while evaluating the significance of each instance segment. During inference, the VAS pain score for the entire video is derived from instance-level predictions. In retrospective analysis using the public UNBC-McMaster dataset, the EDi Pain model demonstrated competitive performance relative to prior studies, achieving strong performance in video-level pain intensity estimation, with a mean absolute error (MAE) of 1.85 and a Pearson correlation coefficient (PCC) of 0.63. Additionally, our model was validated on a prospectively collected dataset of 931 patients from National Taiwan University Hospital, yielding an MAE of 1.48 and a PCC of 0.22. In summary, we developed and validated a novel deep learning-based, instance-appraisable model for pain intensity estimation using facial videos. The EDi Pain model shows promise for real-time application in clinical settings, offering a more objective and dynamic approach to pain assessment.

Keywords: Computer vision; Deep learning; Instance model; Pain estimation; Video.

PubMed Disclaimer

Conflict of interest statement

Declarations. Ethics Approval: The National Taiwan University Hospital (NTUH) Institutional Review Board approved this study (201911054RINA). Consent to Participate: Informed consent was obtained from all participants in the National Taiwan University Hospital study. Consent for Publication: Informed consent was obtained from all participants in the National Taiwan University Hospital study for analyzing and publish their data. The manuscript does not contain any identifiable individual’s images or videos from the National Taiwan University Hospital study. Competing interests: The authors declare no competing interests.

References

    1. Raja SN, Carr DB, Cohen M, Finnerup NB, Flor H, Gibson S, Keefe FJ, Mogil JS, Ringkamp M, Sluka KA, Song XJ, Stevens B, Sullivan MD, Tutelman PR, Ushida T, Vader K. The revised International Association for the Study of Pain definition of pain: concepts, challenges, and compromises. Pain. 2020 Sep 1;161(9):1976-1982. - DOI - PubMed - PMC
    1. Ku NW, Cheng MT, Liew CQ, Chen YC, Sung CW, Ko CH, Lu TC, Huang CH, Tsai CL. Prospective study of pain and patient outcomes in the emergency department: a tale of two pain assessment methods. Scand J Trauma Resusc Emerg Med. 2023 Oct 23;31(1):56. - DOI - PubMed - PMC
    1. Hsu HP, Cheng MT, Lu TC, Chen YC, Liao EC, Sung CW, Liew CQ, Ling DA, Ko CH, Ku NW, Fu LC, Huang CH, Tsai CL. Pain Assessment in the Emergency Department: A Prospective Videotaped Study. West J Emerg Med. 2022 Aug 28;23(5):716-723. - DOI - PubMed - PMC
    1. Gkikas S, Tsiknakis M. Automatic assessment of pain based on deep learning methods: A systematic review. Comput Methods Programs Biomed. 2023 Apr;231:107365 - DOI - PubMed
    1. Prkachin KM, Hammal Z. Computer mediated automatic detection of pain-related behavior: prospect, progress, perils. Front Pain Res (Lausanne). 2021;2:788606. - DOI - PubMed - PMC