查看文章

Nowadays, many of us wear multiple devices capable of acquiring and storing data related to our everyday activities. Since the computing power of mobile battery-operated devices slowly increases and the power optimizations allow for more and more continuous use, these devices are capable of not only monitoring our activity but analyzing the activity as well. Of these devices, the smartwatch is probably the most inconspicuous, and due to its widespread use, we have used accelerometer data gathered from a smartwatch to identify common user activities by using image generated plots and image recognition machine learning. By leveraging the.Net ML.NET machine learning framework we have managed to obtain a decent recognition rate.