Title :
Single Scale Self Quotient Image and PNN for infant pain detection
Author :
Mansor, Muhammad Naufal ; Junoh, Ahmad Kadri ; Ahmed, Amran ; Osman, Muhammad Khusairi
Author_Institution :
Sch. of Mechatron. Eng., Univ. Malaysia Perlis, Arau, Malaysia
Abstract :
Pain is a non-stationary made by babies in reaction to certain circumstances. This infant facial expression can be used to recognize physical or psychology condition of newborn. The goal of this study is to evaluate the performance of illumination levels for infant pain classification. Single Scale Self Quotient Image (SSQ) and discrete Cosine Transform (DCT) features are computed with Probalistic Neural network (PNN) classifier. Eight different performance measurements such as Sensitivity, Specificity, Accuracy, Area under Curve (AUC), Cohen´s kappa (k), Precession, F-Measure and Time Consumption gives a remarkable results with higher than 90%.
Keywords :
discrete cosine transforms; emotion recognition; image classification; neural nets; psychology; DCT features; PNN classifier; SSQ; discrete cosine transform features; illumination levels; infant facial expression; infant pain classification; infant pain detection; newborn babies; performance measurements; physical condition recognize; probabilistic neural network classifier; psychology condition recognize; single scale self quotient image; Accuracy; Conferences; Databases; Discrete cosine transforms; Lighting; Pain; Pediatrics; COPE; Discrete Cosine Transform (DCT); Probalistic Neural network; Single Scale Self Quotient Image (SSQ);
Conference_Titel :
Control System, Computing and Engineering (ICCSCE), 2014 IEEE International Conference on
Print_ISBN :
978-1-4799-5685-2
DOI :
10.1109/ICCSCE.2014.7072779
