Autors: Tsokov, S. A., Lazarova, M. K., AleksievaPetrova, A. P.
Title: An evolutionary approach to the design of convolutional neural networks for human activity recognition
Keywords: Convolutional neural networks; Deep learning; Genetic algorithms; Human activity recognition; Neural network architecture design

Abstract: Automated human activity recognition has a number of applications such as in elderly healthcare monitoring, fitness tracking and in various smart home systems that can adapt to the inhabitants’ behavior. Deep learning using Convolutional Neural Networks (CNNs) is increasingly being used for recognition of human activities. However, the CNNs performance is highly dependent on the network architecture and usually the hyper-parameters are manually selected. Various approaches have been used to automate the design of CNN architectures. The paper proposes an evolutionary based approach for optimizing the architecture of one dimensional CNNs for human activity recognition. The suggested approach is tested on three accelerometric data sets. The experimental results show that the CNNs designed using the evolutionary based approach have better performance on the WISDM Actitracker, Smartphone-Based Recognition of Human Activities and Postural Transitions data set compared to other deep CNNs.

References

    Issue

    Indian Journal of Computer Science and Engineering, vol. 12, issue 2, 2021, India, https://doi.org/10.21817/indjcse/2021/v12i2/211202145

    Copyright Indian Journal of Computer Science and Engineering

    Цитирания (Citation/s):
    1. Fujii, A., Yoshida, K., Shirai, K., Murao, K. (2022). Bento Packaging Activity Recognition with Convolutional LSTM Using Autocorrelation Function and Majority Vote. In: Ahad, M.A.R., Inoue, S., Roggen, D., Fujinami, K. (eds) Sensor- and Video-Based Activity and Behavior Computing. Smart Innovation, Systems and Technologies, vol 291. Springer, Singapore. https://doi.org/10.1007/978-981-19-0361-8_16 - 2022 - в издания, индексирани в Scopus или Web of Science
    2. Saha, A., Rajak, S., Saha, J. et al. A Survey of Machine Learning and Meta-heuristics Approaches for Sensor-based Human Activity Recognition Systems. J Ambient Intell Human Comput (2022). https://doi.org/10.1007/s12652-022-03870-5 - 2022 - в издания, индексирани в Scopus или Web of Science

    Вид: пленарен доклад в международен форум, публикация в реферирано издание, индексирана в Scopus