Pain-attentive network: a deep spatio-temporal attention model for pain estimation
- PDF / 3,104,592 Bytes
- 26 Pages / 439.642 x 666.49 pts Page_size
- 100 Downloads / 156 Views
Pain-attentive network: a deep spatio-temporal attention model for pain estimation Dong Huang1
· Zhaoqiang Xia1 · Joshua Mwesigye1 · Xiaoyi Feng1
Received: 16 October 2019 / Revised: 4 June 2020 / Accepted: 21 July 2020 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract In the video surveillance of medical institutions, pain intensity is a significant clue to the state of patients. Of late, some approaches leverage various spatio-temporal methods to capture the dynamic pain information of videos for accomplishing pain estimation automatically. However, there is still a challenge in the spatio-temporal saliency, which means pain is always reflected in some important regions of informative image frames in a video sequence. To this end, we propose a deep spatio-temporal attention model called as PainAttentive Network (PAN), which pays more attention on the saliency in the extraction of dynamic features. PAN consists of two subnetworks: spatial and temporal subnetwork. Especially, in spatial subnetwork, a proposed spatial attention module is embedded to make the spatial feature extraction more targeted. Also, a devised temporal attention module is inserted in temporal subnetwork, so that the temporal features focus on informative image frames. Extensive experiment results on the UNBC-McMaster Shoulder Pain database show that our proposed PAN achieves compelling performances. In addition, to evaluate the generalization, we report competitive results of our proposed method in the Remote Collaborative and Affective database. Keywords Deep learning · Spatio-temporal model · Attention mechanism · Pain estimation
Zhaoqiang Xia
[email protected] Dong Huang [email protected] Joshua Mwesigye [email protected] Xiaoyi Feng [email protected] 1
School of Electronics and Information, Northwestern Polytechnical University, Xi’an, China
Multimedia Tools and Applications
1 Introduction Pain intensity directly reflects the status of patients in health care, so an accurate and effective method of pain estimation in medical institutions is essential. If the patient is in a state of intense pain but fails to get timely treatment, it will cause a major medical accident. Nowadays, there are two ways to estimate pain intensity for patients in medical institutions, i.e., self-report and observer assessment. The self-report of patients is an easy way to voice painful experience. However, this is not suitable for people who cannot express their pain through speech, e.g., newborns and cerebral palsy patients [61]. Further, the observer’s assessment requires continuous and uninterrupted care by family members or medical staffs, which is extremely inefficient. Therefore, it is necessary to devise some solutions of estimating pain intensity automatically, which greatly improves the efficiency of the medical institution and the psychological comfort of patients. Earlier, some researchers were inspired by the basic approaches of facial expressions recognition and applied these approaches to pain est
Data Loading...
