International Journal of Computer Techniques Volume 12 Issue 3 | Sign2Text: Bridging communication for Deaf and Non-Speaking Individuals

Sign2Text: Bridging Communication for Deaf and Non-Speaking Individuals

Sign2Text: Bridging Communication for Deaf and Non-Speaking Individuals

Authors: Deepali Kumari, Ram Kumar Sharma

Abstract

Sign language enables communication for individuals with speech impairments. Our system integrates CNN-RNN deep learning models with HMM-based speech synthesis to convert sign gestures into text and speech, enhancing accessibility.

Keywords

HMM, ASL, BSL, CNN, RNN, Gesture recognition, Deep learning, Text-to-Speech

Conclusion

The proposed system successfully recognizes sign gestures and translates them into speech. Future work includes improving real-time accuracy using Transformer models and refining dataset annotation techniques.

References

1. L.O. Chua, “CNN: A Vision of Complexity,” Int. J. Bifurcation Chaos, vol. 7, no. 10, 1997.

2. S. Mascarenhas, M. Agarwal, “Comparison between VGG16, VGG19, and ResNet50 architectures,” CENTCON 2021.

Post Comment