Transforming Deep Learning Models for Resource-Efficient Activity Recognition on Mobile Devices

Mobile and wearable sensor technologies have gradually extended their usability into a wide range of applications, from well-being to healthcare. The amount of collected data can quickly become immense to be processed. These time and resource-consuming computations require efficient methods of classification and analysis, where deep learning is a promising technique. However, it is challenging to train and run deep learning algorithms on mobile devices due to resource constraints, such as limited battery power, memory, and computation units. In this paper, we have focused on evaluating the performance of four different deep architectures when optimized with the Tensorflow Lite platform to be deployed on mobile devices in the field of human activity recognition. We have used two datasets from the literature (WISDM and MobiAct) and trained the deep learning algorithms. We have compared the performance of the original models in terms of model accuracy, model size, and resource usages, such as CPU, memory, and energy usage, with their optimized versions. As a result of the experiments, we observe that the model sizes and resource consumption were significantly reduced when the models are optimized compared to the original models.