EURASIP Journal on Image and Video Processing

Table 2 Comparison results of the proposed with other state-of-the-art action recognition methods

From: Consistent constraint-based video-level learning for action recognition

Training method	Method	Accuracy (%)
Clip level	C3D [2]	51.52
	Geometry [32]	55.2
	CD-UAR [33]	42.5
	3D-ShuffleNetV2 [34]	56.52
	MASN [23]	53.44
Video level	Ours	58.76

Back to article page