International Journal of Advances in Electrical Engineering
2024, Vol. 5, Issue 1, Part A
Human activity detection using deep learning approaches
Author(s): Harish Chandra Sati and Tanish
Abstract: One of the most crucial computer vision jobs is the recognition of human activity, which has proven beneficial in a variety of industries like security, healthcare, and athletic training. Several methods have been investigated to do this task, some of which make use of sensor data and others use video data. In this paper, we explore two deep learning-based methods to recognize human activities in videos: convolutional long short-term memory and single frame convolutional neural networks (CNNs). It is preferable to use a convolutional neural network-based approach since CNNs can automatically extract features and long short-term memory networks are excellent for working with sequence data, like video. The two models were trained and assessed using an action recognition benchmark. On the UCF50 benchmark action recognition dataset as well as a second dataset generated specifically for the experiment, the two models were trained and assessed. Despite the good accuracy of both models, the single frame.
DOI: 10.22271/27084574.2024.v5.i1a.46
Pages: 01-09 | Views: 358 | Downloads: 222
Download Full Article: Click Here
How to cite this article:
Harish Chandra Sati, Tanish. Human activity detection using deep learning approaches. Int J Adv Electr Eng 2024;5(1):01-09. DOI: 10.22271/27084574.2024.v5.i1a.46
