1 paper
Ke Xing, Zhilong Wang, Zheng Lian +8
Existing affective benchmarks largely rely on fixed recordings, where observation conditions are predetermined before inference. Consequently, they mainly evaluate passive affect r…