Authors:
R. Regin, S. Thkashinraj, J. Koushik Raghav, S. Jude Roosevelt, G. Padmapriya, Chunhua Deming

Addresses:
Department of Computer Science and Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Chemistry, Bharath Institute of Higher Education and Research, Chennai, Tamil Nadu, India. Information Technology Discipline, NUS Graduate School (NUSGS), National University of Singapore, Queenstown, Singapore. 

Abstract:

The posture detection application is a cutting-edge tool leveraging TensorFlow Lite and OpenCV to analyze real-time human movement. Designed for motion tracking, occupational safety, rehabilitation, and interactive technology applications, the system provides accurate predictions by analyzing anatomical landmarks. It processes live video and still images, providing high reliability and accuracy essential for sports training, ergonomic evaluation, and healthcare. The application provides instant feedback on posture and technique, allowing users to improve their performance dynamically. It issues alerts if key indicators are not detected within a specified time, ensuring timely intervention in critical scenarios such as security and elderly care. Features include intuitive controls, a confidence threshold slider to adjust accuracy, and support for uploading still images. The user interface visually represents key landmarks, facilitates data interpretation, and features voice commands for improved usability. Combining advanced machine learning and user-centric design, the system supports real-time monitoring and analysis, making it versatile and effective for improving performance, ensuring security, and delivering interactive experiences in various environments.

Keywords: Human Pose Estimation; MoveNet Model; Real-Time System; Keypoint Detection; Inactivity Monitoring; Occupational Safety; Multi-Person Detection; TensorFlow Lite; Graphical User Interface (GUI).

Received on: 17/04/2024, Revised on: 26/06/2024, Accepted on: 17/08/2024, Published on: 09/09/2024

DOI: 10.69888/FTSIN.2024.000287

FMDB Transactions on Sustainable Intelligent Networks, 2024 Vol. 1 No. 3, Pages: 178-189

• Views : 38
• Downloads : 8

Download PDF