https://github.com/mrfoxak/evaluate-lip-reading-using-deep-learning-techniques.

This paper explores Silent Sound Technology, focusing on its potential to enhance communication in noisy environments through lip-reading and deep learning, with applications in hearing aids and security.
https://github.com/mrfoxak/evaluate-lip-reading-using-deep-learning-techniques.

bi-lstm cnn cuda deep-learning image-processing lstm machine-learning mathematics neural-networks ovencv python research-paper sklearn tensorflow

Last synced: 10 months ago
JSON representation

This paper explores Silent Sound Technology, focusing on its potential to enhance communication in noisy environments through lip-reading and deep learning, with applications in hearing aids and security.

• Host: GitHub
• URL: https://github.com/mrfoxak/evaluate-lip-reading-using-deep-learning-techniques.
• Owner: MrfoxAK
• Created: 2024-08-21T16:04:11.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-08-24T16:04:15.000Z (almost 2 years ago)
• Last Synced: 2025-04-15T01:12:01.813Z (about 1 year ago)
• Topics: bi-lstm, cnn, cuda, deep-learning, image-processing, lstm, machine-learning, mathematics, neural-networks, ovencv, python, research-paper, sklearn, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 2.1 MB
• Stars: 3
• Watchers: 1
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
## Evaluate-Lip-reading-using-Deep-Learning-Techniques.

This paper explores Silent Sound Technology, focusing on its potential to enhance communication in noisy environments through lip-reading and deep learning, with applications in hearing aids and security.

## Dataset:

LRW, GRID, LRS2, LRS3-TED, VoxCeleb2, CAS-VSR-W1k(LRW-1000

### Contact

For any questions or feedback, please contact 

[Debojyoti Bhuinya](https://debojyotibhuinya-portfolio.netlify.app/),

[Subhamay Ganguly](https://www.linkedin.com/in/subhamay-ganguly-526972248/),

[Akash Das](https://www.linkedin.com/in/akash-das-80b356230/),

## Proposed Methodology

In this section, we describe the methodology employed for preprocessing video data,

aligning it with textual content, and training a neural network model for a specific

task. The proposed methodology encompasses data loading, data preprocessing, and

model training.

![Screenshot 2024-08-21 220201](https://github.com/user-attachments/assets/1c1f007b-1e43-403d-86ba-867477ea096d)

## Convolutional Neural Network:

![Screenshot 2024-08-21 220304](https://github.com/user-attachments/assets/2eb27458-5fa5-4d4a-b3e4-1da380a150d1)

## Bi-Directional LSTM architecture

![Screenshot 2024-08-21 220423](https://github.com/user-attachments/assets/83b493c3-d91a-4699-91e8-c18532958f18)

## Total Architecture

![update](https://github.com/user-attachments/assets/b12774e1-eb09-4e14-b720-dae62bc78189)

## Performence Evaluation

![acuracy](https://github.com/user-attachments/assets/2a6152a5-cce2-468e-b136-9681306b25d0)