AffectNet
burak yılmaz
홈페이지 · 논문 323편
AffectNet is a large facial expression dataset with around 0.4 million images manually labeled for the presence of eight (neutral, happy, angry, sad, fear, surprise, disgust, contempt) facial expressions along with the intensity of valence and arousal. Source: Landmark Guidance Independent Spatio-channel Attention and Complementary Context Information based Facial Expression Recognition Image Source: http://mohammadmahoor.com/affectnet/
Images EnglishArabic