Attention-based CNN-LSTM architecture for smartphone IMU-based human activity recognition in intelligent mobile monitoring systems
An attention-based CNN–LSTM–Attention architecture for smartphone-based HAR using accelerometer signals is proposed, confirming that convolutional feature extraction, recurrent temporal modeling, and attention-based weighting improve the reliability of accelerometer-based HAR while maintaining practical computational e...
M. Turdalyuly, Nazerke Baisholan, Sergiy Gnatyuk et al.
· 1 citation