Real-Time Human Action Recognition and Alert System

DOI

10.2991/978-94-6239-713-2_47How to use a DOI?

Keywords

Human Action Recognition; MediaPipe; Random Forest Classifier; Arduino Uno; GSM Module; Smart Surveillance

Abstract

Human Action Recognition (HAR) is a major field of study in both smart surveillance systems and computer vision. Human activities such as falls or fights are of great concern due to the need for immediate contact to avert injuries and preserve public safety. This research proposes a real-time HAR and emergency alerting system using IoT hardware and Machine Learning technology. The proposed HAR system consists of a camera and MediaPipe Pose for recognizing human body landmarks in real time. The detection method is hybrid and uses rule-based logic to recognize basic activities (e.g., standing, sitting, walking, reading, clapping, and falling) and employs a Machine Learning model to recognize fighting activity only. A Random Forest Classifier is used to model pose landmark features across a variety of actions, providing multidimensional information about the detected activity. When a fall or fight is detected, an alert signal is sent from the detection system (camera/media pipe) to the Arduino Uno via a Serial Communication interface. The Arduino Uno activates the buzzer, displays warning messages on an LCD screen, and sends out an emergency SMS via a GSM module. Based on the experimental results, the proposed system achieves 99.7% accuracy and operates effectively in real time. The accuracy was calculated by testing on a collected dataset consisting of 35,000 pose landmark samples, split into an 80:20 train-test ratio. Finally, the proposed HAR and notification system is a practical and computationally efficient solution for safety-based surveillance applications.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Aryan Kumar
AU  - Dhruv Swami
AU  - R. Kavitha
PY  - 2026
DA  - 2026/06/25
TI  - Real-Time Human Action Recognition and Alert System
BT  - Proceedings of the International Conference on Advances in Computing Technology and Artificial Intelligence (COMPUTATIA 2026)
PB  - Atlantis Press
SP  - 634
EP  - 640
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6239-713-2_47
DO  - 10.2991/978-94-6239-713-2_47
ID  - Kumar2026
ER  -