Ambulatory physical activity representation and classification using spectral distances approach

Methods for monitoring the human physical activity are recently investigated in order to assess the health status of the individuals and thus promote a healthier lifestyle. This paper proposes `dist-colorimetrics´, a methodology that aims to represent and classify ambulatory activities based on the spectral distances measures. A data collection platform including four accelerometer sensors mounted on the chest, ankle, wrist and hip, is used to record five activities: running, walking, cycling, resting and car riding. The proposed approach converts raw acceleration data into relevant spectral distances parameters. A 2D colored illustration of these parameters provides efficient visual representation as to the similarity and the variation among activities. For a further validation in terms of recognition performance, the `dist-colorimetrics´ model was trained and tested by implementing three classification techniques, namely the Naïve Bayes, the K-nearest neighbors and the decision tree. The results showed that the system reached up to 98.12% of overall recognition accuracy. With further improvement in the modeling of each activity, we have reason to believe that the spectral distances are a promising approach to distinguish between different physical activities.