Human Depth Sensors-Based Activity Recognition Using Spatiotemporal Features and Hidden Markov Model for Smart Environments

Table 2

Recognition results of proposed HAR system.

Actions	Accuracy
High arm wave	93.7
Horizontal arm wave	90.4
Hammer	89.2
Hand catch	95.5
Forward punch	91.1
High throw	89.7
Drawing X	93.4
Drawing tick	97.1
Drawing circle	92.3
Hand clap	89.7
Two-hand wave	94.5
Side boxing	92.4
Bending	86.8
Forward kicking	91.3
Side kicking	98.6
Jogging	96.7
Tennis swing	87.1
Tennis serve	86.9
Golf swing	97.8
Pickup and throw	94.3
Mean recognition accuracy = 92.4