The increasingly acceptance of Internet of Things (IoT) devices adversely contribute to the accumulation of massive amounts of data which challenges for the adoption of techniques capable to handle it. This paper first presents overall points of deep learning, and IoT principles. After that, this paper uses the human recognition activity scenario to evaluate two DL models: the Convolutional Neural Network (CNN), and the Recurrent Neural Network. At last, a benchmark with the state-of-the-art is presented. The main findings evidenced the suitability of the proposed model; the CNN performed a mean accuracy rate of 93%, and therefore it is likely to be embedded in an IoT device. There is room for improvements, namely, the ability to recognize additional human activities, and to include more testing scenarios.
CITATION STYLE
Mendes, T., & Pombo, N. (2023). Evaluation of Deep Learning Techniques in Human Activity Recognition. In Lecture Notes in Networks and Systems (Vol. 542 LNNS, pp. 114–123). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-16072-1_8
Mendeley helps you to discover research relevant for your work.