mabu65 / ppg-pattern-recognition

Pattern recognition in PPG signals using only a limited number of labeled examples. We train a network to separate clean segments from noise and motion artifacts.

Semi-supervised pattern recognition in PPG signals

Please refer to the blog post for more information on this project if you are new.

The purpose of this tutorial is to present self-training as a viable semi-supervised approach for regular pattern recognition in physiological signals, specifically using PPG or pulse oximetry, with only a few pre-labeled examples. The trained model can be applied during the signal pre-processing stage to distinguish relatively clean segments from ones affected by substantial noise and/or motion artifacts. It can achieve an observed accuracy as high as 94.7% using a limited initial training subset of 500 labels, and shows a performance increase compared to a 91% accuracy with a fully supervised model trained against the entire available labeled subset.

Code and usage

Main dependencies: Tensorflow/Theano, Keras, scikit-learn.

Clone the repo and call main.py with no arguments to train the model and visualize plots of the confusion matrix, loss over epochs, and a few example segments demonstrating the classifier results, as shown below. Note: the code was written and tested in Python 3.

*Comparison of two types of labeled signals from the dataset, showing the original signal in blue, and a bandpass filtered version in red. Above: segment with high quality signal content, labeled positive. Below: segment content is random and should be rejected, labeled negative.*