A Video Classification Method Based on Spatiotemporal Detail Attention and Feature Fusion

Table 2

Exploration of spatial-temporal consistency and multimodal fusion of kinetics 400.

	Top1	Top5
VCM-SDD No_STC No_MFF	71.7	90.9
VCM-SDD STC No_MFF	73.9	91.4
VCM-SDD No_STC MFF	78.5	93.6
VCM-SDD STC MFF	80.1	94.4
TSN [1]	71.3	91.5
TSM [2]	75.1	91.8