PPG Dataset of 219 people is given, use any of the DNNs to classify the 2D images of the spectograms acquired from the given data.
Convert the signal obtained from samples into 2-D spectrograms and classify the obtained images using pre-trained CNN models (like Alexnet, Resnet, Mobilenet etc.).
The CNN (Convolutional Neural Network) used in this project is called AlexNet.
- Given PPG Signal -
- PPG Signal Processing -
- Sectioning of PPG signal for data balancing -
- Spectogram of Sectioned Signal -
- Normal Condition Spectogram -
- Prehyptersion Condition Spectogram -
- Stage-1 Hypertension Spectogram -
- Stage-2 Hypertension Spectogram -
The spectogram dataset is then used for training and testing of the CNN model. The model got overfitted since the data size was not adequate. The dataset was custom, and still being updated.
* Results -