Sign Language Recognition using Mediapipe and RNN Models(LSTMand GRU)
Title: Sign Language Recognition Using Mediapipe and RNN Models (LSTM and GRU)
Permalink: sign-language-recognition-mediapipe-rnn-lstm-gru
Description: This research integrates the Mediapipe framework with Recurrent Neural Networks (RNNs), including Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU), to enhance the accuracy of sign language recognition. The study compares both models, demonstrating GRU’s superior performance with an accuracy of 96%-99%.
Focus Keywords: Sign Language Recognition, Mediapipe, LSTM, GRU, Recurrent Neural Networks, Hand Gestures, AI-driven communication
International Journal of Engineering and Techniques – Volume 10 Issue 2, March 2024
www.ijetjournal.org
ISSN: 2395-1303
M. Kishore Babu1, M. Venkata Karthik Reddy2, M. Lavanya3, N. Sai Raghu Vardhan4, M. Vijay Kumar5
1Assistant Professor, Department of CSE, Vasireddy Venkatadri Institute of Technology (Autonomous), Guntur, AP.
2,3,4,5UG Students, Department of CSE, Vasireddy Venkatadri Institute of Technology (Autonomous), Guntur, AP.
Emails: kislatha@gmail.com, karthikmedagam@gmail.com, lavanyaml.1702@gmail.com, sairaghu241@gmail.com, vjaymogili@gmail.com
Abstract
This project aims to bridge the communication gap between regular individuals and the differently abled by utilizing the Mediapipe framework and RNN models like LSTM and GRU. The integration of Mediapipe, a real-time hand-tracking library, with Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU) enhances the accuracy and efficiency of sign language recognition. The system processes recorded video samples, extracts key hand and facial landmarks using Mediapipe, and applies RNN-based models for classification. The study found that GRU models outperform LSTM models in sign language recognition accuracy, achieving between 96%-99% compared to LSTM’s 90%-94% accuracy.
Keywords
Mediapipe, LSTM, GRU, RNN, Hand Gestures, Sign Language, AI-driven Communication
How to Cite
M. Kishore Babu, M. Venkata Karthik Reddy, M. Lavanya, N. Sai Raghu Vardhan, M. Vijay Kumar, “Sign Language Recognition Using Mediapipe and RNN Models (LSTM and GRU),” International Journal of Engineering and Techniques, Volume 10, Issue 2, March 2024.
Post Comment