Designing an Efficient System for Emotion Recognition Using CNN

Table 8

Result on FER2013.

Method	Accuracy (%)	Gain (%)
AlMarri’s method [15]	30.19	35.7
Mollahosseini et al.’s method [12]	61.10	4.79
Haque and Valles’ method [11]	63.10	2.79
Zeng et al.’s method [43]	61.86	4.03
Xie et al.’s method [14]	62.99	2.90
Tümen et al.’s method [13]	57.10	8.79
Github et al.’s method [16, 17]	65.34	0.55
Proposed method	65.89