This work is cited from https://ieeexplore.ieee.org/document/9156120
The tutorial is in 'CNN_Recurrent.ipynb'. We also have python version for the project, and the file is 'CNN_Recurrent.py'. There are two additional function in CNN_Recurrent.py
- Add 'Changable hidden layers' in Class RCNN, which means you can change the model sturcture by changing 'self.hidden layer'
- Add 'plot result', which can show the loss vs epoch plot, and save it in ./data/WristSense/CNNs_HAR_and_HR.
And the datasets are in the ./data/WristSense/CNNs_HAR_and_HR/Data/wrist The data was download from https://www.physionet.org/content/wrist/1.0.0/