Pain Assessment in Infants: Towards Spotting the Pain Expression Based on the Facial Strain
pain, pediatrics, strain, video sequences, classification algorithms, support vector machines, accuracy, SVM classifiers, pain assessment, pain expression, facial strain, facial tissue distortion, pain indicator, video-sequence, facial expression classification, k-nearest-neighbor classifier, KNN classifier, support vector machine classifier
Digital Object Identifier (DOI)
We report novel results of utilizing infant facial tissue distortion as a pain indicator in video-sequences of ten infants based on analysis of facial strain. Facial strain, which is used as the main feature for classification, is generated for each facial expression and then used to train two classifiers, k Nearest-Neighbors (KNN) and support vector machine (SVM) to classify infants' expressions into two categories, pain and no-pain. The accuracy of binary classification for KNN and SVM was 96% and 94% respectively, based on ten video sequences. The results of this study are encouraging; they indicate that assessing pain based on facial displays is a promising area of investigation, and open new directions for future work.
Was this content written or created while at USF?
Citation / Publisher Attribution
2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG), Ljubljana, p. 1-5.
Scholar Commons Citation
Zamzmi, Ghada; Ruiz, Gabriel; Goldgof, Dmitry; Kasturi, Rangachar; Sun, Yu; and Ashmeade, Terri, "Pain Assessment in Infants: Towards Spotting the Pain Expression Based on the Facial Strain" (2015). Computer Science and Engineering Faculty Publications. 78.