Table 3. Performance comparison with literature on the JAFFE dataset.

Methods	Evaluation strategy	Accuracy (%)
Pre-processing + CNN (Lopes et al., 2017)	10-fold	37.36
IFSL (SVM) (Yan et al., 2020),	10-fold	88.2
AlexNet + FC6 + LDA (Fei et al., 2020)	5-fold	95
Salient facial patches based FER (Happy & Routray, 2014)	10-fold	91.80
LiveEmoNet (Podder, Bhattacharya & Majumdar, 2022b)	10-fold	96.36
Hybrid Methodology (Khan et al., 2019)	10-fold	92.16
Dense_FaceLiveNet (Hung, Lin & Lai, 2019)	5-fold	90.97 (±3.95)
MobileNet with Softmax + Center loss (Miao et al., 2019)	5-fold	95.24
Proposed method	10-fold	96.83 (±3.15)