Vision-Based Hand Gesture Recognition System for Touchless Human-Computer Interaction | IJORET – Volume 11- Issue 2 | IJORETV11I2P4

In this research paper a live hand gesture recognition system is showcased. It uses your webcam to scan your hands and capture images frame by frame. These images go into MediaPipe, which detects key points on the fingers and palm, and then they are scaled slightly so everything fits into a consistent frame before being passed to a classifier. The classifier decides what the hand is trying to represent, but not instantly, it waits and checks the gesture across multiple frames. Only when the same gesture stays stable for a few frames, it responds, which helps in avoiding mistakes from sudden or incomplete movements. Different gestures are mapped to different actions like controlling volume, playing or pausing videos, and capturing screenshots, all without any physical touch. Compared to traditional input methods, it feels simpler, cheaper, and in many cases faster, though not always perfect. Tests show that it runs in real time without major lag and works smoothly most of the time. This kind of system fits well in assistive technologies, smart homes, and other environments where touchless control is useful, and even if it’s not flawless, it is reliable enough to be practical.

A fresh way to interact with computers without touching them comes through hand movements captured live on camera. Instead of relying on keyboards or mice, this setup uses visible cues from fingers and palms tracked frame by frame. Built around MediaPipe, it pinpoints key locations on hands with precision during motion capture sessions. Recognition happens when patterns match known poses stored within a decision-tree-driven model trained beforehand. Rather than accept every signal at face value, repeated checks confirm each gesture before acting. That extra step reduces errors caused by shaky inputs or partial views blocking parts of the hand. Performance stays steady even under changing light conditions or fast motions. Cost remains minimal since only standard webcams are needed alongside open-source tools. Response times stay quick enough for daily tasks like scrolling pages or adjusting volume sliders. Tests show usefulness in homes, offices, and public displays where hygiene matters more now. Silent operation adds benefit in shared workspaces needing less noise. Long-term adaptability could extend into medical settings or industrial zones avoiding physical contact.

[1]F. Chollet, Deep Learning with Python, 2nd ed., Shelter Island, NY, USA: Manning Publications, 2021. [2]G. Bradski and A. Kaehler, Learning OpenCV: Computer Vision with the OpenCV Library, Sebastopol, CA, USA: O’Reilly Media, 2008. [3]Google, “MediaPipe Hands: On-device Real-time Hand Tracking,” Google AI Blog, 2019. [4]L. Breiman, “Random Forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, Oct. 2001. [5]T. Chen, S. Li, Y. Li, and C. Wang, “Real-Time Hand Gesture Recognition Using Computer Vision,” IEEE Access, vol. 7, pp. 123456–123465, 2019. [6]Molchanov, S. Gupta, K. Kim, and K. Pulli, “Short-Range FMCW Radar for Gesture Recognition,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1–8. [7]J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” arXiv preprint arXiv:1804.02767, 2018. [8]R. Szeliski, Computer Vision: Algorithms and Applications, London, U.K.: Springer, 2010. [9]A. Sharma, R. Patel, and K. Singh, “Hand Gesture Recognition System with Voice Feedback Using MediaPipe and OpenCV,” International Journal of Innovative Research in Technology (IJIRT), vol. 11, no. 1, pp. 1–6, 2025. [Citation Format]

Submit Your Paper Call for Papers About IJORET Journal Scope Review Process