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

Table 4

Recognition results of proposed HAR system.

Activities	Accuracy
Drink	95.4
Eat	87.8
Read book	96.3
Call phone	89.1
Write	96.7
Use laptop	90.5
Use cleaner	100
Cheer up	88.5
Sit still	94.6
Toss paper	94.2
Play game	86.3
Lay down	98.7
Walk	95.3
Play guitar	88.2
Stand up	96.5
Sit down	93.6
Mean recognition accuracy = 93.2