In this research, the potential of PPG signals was examined in recognition of 14 affective states. The PPG signs of DEAP database had been chosen. The novel PPG descriptors, ECWS, was presented. The recognition was done applying PNN. Recently, in feelings recognition literature, PNN have been successfully applied to physiological alerts (Goshvarpour et al., 2017a, 2017b). In addition , the effect of changing the sigma variable in the efficiency of the classifier was examined.
Because shown in Figure four, the maximum accuracy rate of 100% was achieved using sigma 0. 001 to 0. 301. Discounting the sigma parameter, the classification rates were in the selection of 74. thirty seven to 100% for classifying 14 emotion classes.
Table a couple of shows acceptable results in the proposed construction with the research on programmed emotion recognition using PPG signals. Verhoef et approach. (2009) utilized PPG and GSR morphological indices to create emotion recognition. For GSR, the number of reactions means extravagance, mean increasing time, as well as the mean strength of the answers was removed. For PPG, heart rate variability (HRV), the mean interbeat interval (IBI), the standard deviation of the IBI, and the imply amplitude from the IBI were computed. By using a static Bayesian network, the highest performance of 60% was reported inside the classification of 7 emotion classes. Koelstra et al. (2012) examined the multi-modal DEAP signals of 32 subject matter to classify valence and arousal categories. They will fed the spectral power asymmetry in the four ELEKTROENZEPHALOGRAFIE bands into the Fisher and NaÃ¯ve Bayes. The valence and excitement levels based feelings were labeled with the charge of 57. 6 and 62%, respectively. Park ainsi que al. (2013) used ST and PPG signals of 5 volunteers to classify delight and sadness. A time span between two successive PPG peaks and ST exuberance were appended by SVM. It has been reported that the accuracy and reliability rate of ST was 89. 29%, the recognition rate of PPG was 63. 67%, and a put together feature ended in the highest reputation rate of 92. 41%. An feelings recognition structure was suggested by Li et ing. (2014) using statistical features of ECG, GSR, and PPG. MLP could recognize four categories of emotion with the greatest rate of 78. 06%. Recently, Khan and Lewo (2016) used a ready-made platform pertaining to PPG and GSR. That they tested two classifiers including decision tree (J48) and IBK to categorize almost eight classes of emotion. The maximum accuracy involved 92%.