# Indian Sign Language Recognition Using Mediapipe Holistic

**Authors:** Dr. Velmathi G, Kaushal Goyal  
**arXiv:** [2304.10256](https://arxiv.org/abs/2304.10256)  
**Date:** 2023-04-20  
**Summary page:** https://www.summarizepaper.com/en/paper/2304.10256v1/

## Key points

- Deaf individuals face communication challenges due to their inability to hear
- Technology can help overcome these barriers by providing alternative communication channels
- Sign language recognition (SLR) facilitates communication between deaf and hearing individuals
- The study proposes a robust SLR system for Indian Sign Language conversion
- Convolutional Neural Network (CNN) performs better in recognizing static sign language, while Long Short-Term Memory (LSTM) excels in gesture sign language phrases and sentences
- Text-to-sign language paradigm enhances communication skills of the deaf population
- Not all individuals with hearing impairments are proficient in reading or writing text
- A text-to-sign language paradigm enables understanding of text-based information and participation in various settings
- Technology plays a significant role in overcoming communication barriers for deaf individuals.

## Summary

Deaf individuals face numerous communication challenges in their daily lives due to their inability to hear, making it difficult for them to communicate with those who do not understand sign language. These obstacles extend to educational, occupational, and social contexts. However, technology can play a crucial role in overcoming these barriers by providing alternative communication channels. One such technology that facilitates communication between deaf and hearing individuals is sign language recognition (SLR). In this study titled "Indian Sign Language Recognition Using Mediapipe Holistic," Dr. Velmathi G and Kaushal Goyal propose the development of a robust SLR system specifically designed to convert Indian Sign Language into text or speech. The researchers evaluate the proposed system by comparing Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) models. Since there are both static and gesture sign languages, a robust model is required to distinguish between them effectively. The study reveals that the CNN model performs better in recognizing letters and characters of static sign language, while the LSTM model outperforms CNN in monitoring hands, faces, and pose for gesture sign language phrases and sentences. The creation of a text-to-sign language paradigm is essential as it enhances the communication skills of the sign language-dependent deaf and hard-of-hearing population. While sign-to-text translation addresses one side of communication, not all individuals with hearing impairments are proficient in reading or writing text. Some may struggle with written language comprehension due to educational or literacy issues. Therefore, a text-to-sign language paradigm would enable them to understand text-based information and actively participate in various social, educational, and professional settings. Overall, this research highlights the significance of technology in overcoming communication barriers faced by deaf individuals.