Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy

Roshani Rajesh khobragade; Rutul Ganthade; Aastha Nitnaware; Palak Janbandhu; Amol Bamrotwar; Dr. Archna Potnurwar

Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy

December 2025

Vol-11, Issue-6

Paper ID: 27836

ISSN: 2395-4396

Downloads: 0

Abstract & Details

Research Area

Information Technology / Computer Engineering / Machine Learning

Keywords

Hand gesture recognition Real-time mobile gesture detection MediaPipe Hands TensorFlow Lite Keras deep learning Computer vision OpenCV NumPy Human–computer interaction Mobile deployment Landmark extraction Dynamic gesture recognition

Abstract

Hand gesture recognition has emerged as an intuitive and natural mode of human–computer interaction, especially for mobile and embedded platforms where traditional input mechanisms are limited. This paper presents a comprehensive framework for real-time hand gesture recognition designed specifically for mobile devices. The system integrates MediaPipe for lightweight and accurate hand-landmark extraction, Keras/TensorFlow for model training, and TensorFlow Lite for optimized on-device inference. OpenCV and NumPy are utilized for image preprocessing, frame handling, and feature engineering. A custom dataset of static and dynamic hand gestures was collected under multiple lighting and background conditions to evaluate real-world performance. The proposed classifier achieved an accuracy of 96.8% for static gestures and 91.4% for dynamic sequences, with an average end-to-end latency of ~20 ms on mid-range Android devices. These results demonstrate the feasibility of deploying low-latency, resource-efficient gesture recognition models on mobile platforms. The study further discusses challenges such as occlusion, lighting variations, multi-hand scenarios, and offers recommendations for improving robustness and expanding gesture vocabularies. The findings highlight the potential of real-time gesture-based interfaces in applications such as accessibility, AR/VR interaction, IoT control, and touchless mobile navigation

License

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Author Information

#	Name	Institute / Affiliation
1	Roshani Rajesh khobragade	Priyadarshini College of Engineering, Hingna, Nagpur
2	Rutul Ganthade	Priyadarshini College of Engineering, Hingna, Nagpur
3	Aastha Nitnaware	Priyadarshini College of Engineering, Hingna, Nagpur
4	Palak Janbandhu	Priyadarshini College of Engineering, Hingna, Nagpur
5	Amol Bamrotwar	Priyadarshini College of Engineering, Hingna, Nagpur
6	Dr. Archna Potnurwar	Priyadarshini College of Engineering, Hingna, Nagpur

How to Cite

Use the following formats to cite this article in your research.

APA Style

khobragade, Roshani Rajesh, Ganthade, Rutul, Nitnaware, Aastha, Janbandhu, Palak, Bamrotwar, Amol, & Potnurwar, Dr. Archna (2025). Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy. International Journal of Advance Research and Innovative Ideas In Education, 11(6), 1482-1483.

MLA Style

khobragade, Roshani Rajesh, et al. "Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy." International Journal of Advance Research and Innovative Ideas In Education, vol. 11, no. 6, 2025, pp. 1482-1483.

IEEE Style

Roshani Rajesh khobragade, Rutul Ganthade, Aastha Nitnaware, Palak Janbandhu, Amol Bamrotwar, and Dr. Archna Potnurwar, "Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy," International Journal of Advance Research and Innovative Ideas In Education, vol. 11, no. 6, pp. 1482-1483, 2025.

Vancouver Style

khobragade Roshani Rajesh, Ganthade Rutul, Nitnaware Aastha, Janbandhu Palak, Bamrotwar Amol, Potnurwar Dr. Archna. Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy. International Journal of Advance Research and Innovative Ideas In Education. 2025;11(6):1482-1483.

Harvard Style

khobragade, Roshani Rajesh, Ganthade, Rutul, Nitnaware, Aastha, Janbandhu, Palak, Bamrotwar, Amol, & Potnurwar, Dr. Archna (2025) 'Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy', International Journal of Advance Research and Innovative Ideas In Education, 11(6), pp. 1482-1483.

Chicago Style

khobragade, Roshani Rajesh, et al. "Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy." International Journal of Advance Research and Innovative Ideas In Education 11, no. 6 (2025): 1482-1483.

Turabian Style

khobragade, Roshani Rajesh, et al. "Comprehensive Framework for Real-Time Hand Gesture Recognition on Mobile Platforms using Machine Learning, TensorFlow Lite, Keras, MediaPipe, OpenCV and NumPy." International Journal of Advance Research and Innovative Ideas In Education 11, no. 6 (2025): 1482-1483.