preview

Estimating Pain Intensity Analysis

Decent Essays

\subsection{Pain Intensity Estimation}
Estimating the intensity of the detected pain might provide better pain assessment and lead to better pain management. Several pain recognition methods were extended to include pain intensity estimation. For example, Gholami et al. \cite{5415598} presented a method to estimate infants’ pain intensity using RVM, which is an extension of SVM (See Table II, 3rd row). RVM is a sparse Bayesian model that provides posterior probabilities (i.e., uncertainty) for class memberships through Bayesian inference. Unlike SVM, RVM classifier outputs the probabilities of the class memberships or labels. Gholami et al. used RVM uncertainty for each class membership to estimate infants’ pain intensity. For validation, RVM posterior probabilities were compared with the results of estimating pain intensity by experts and non-expert observers. The agreement between RVM and human observers, measured using kappa coefficient, was 0.48 for experts and 0.52 for non-experts.

Hammal et al. \cite{Hammal:2012:ADP:2388676.2388688} described a method to estimate pain intensities for 25 subjects with orthopedic injury. The proposed method trained four SVM classifiers separately to automatically assess four levels of pain. The reliability of …show more content…

presented a method \cite{gruss2015pain} to estimate four levels of pain using SVM. Facial expression and biopotentials signals were recorded under four levels of pain (T1 to T4) as discussed in Section 4.2. The recorded signals were analyzed to extract complex mathematical features. These features were then used to build SVM that is trained with 75\% of the data and tested on 25\% of data. The proposed method achieved 76.00\% (sensitivity) and 82.59\% (specificity) for baseline vs T1, 80.00\% (sensitivity) and 82.59\% (specificity) for baseline vs T2, 84.71\% (sensitivity) and 85.18\% (specificity) for baseline vs T3, and 92.24\%, (sensitivity) and 89.65\% (specificity) for baseline vs

Get Access