Human Behavior Recognition Method Based on Edge Intelligence

Table 2

Edge node and cloud fusion recognition results (accuracy rate).

	Cross-subject	Cross-view
	81.2	87.1
	82.6	88.0
	80.7	86.9
	81.9	87.4
	82.5	87.7
Single cloud	82.1	87.9
Fusion	83.9	89.7

In the two verification standards, the recognition accuracy rate after fusion is increased, which proves that the advantages of edge-cloud collaboration have made the model recognition accuracy rate steadily increase by more than 2.2%.