https://github.com/adrianocleao/talking-hands

Talking Hands is a project that uses MediaPipe and TensorFlow/Keras for real-time recognition of Brazilian Sign Language (LIBRAS). MediaPipe captures hand, facial, and body landmarks, which are processed by a TensorFlow-based neural network to transcribe signs into text and convert them into Portuguese audio.
https://github.com/adrianocleao/talking-hands

computer-vision deep-learning keras libras machine-learning mediapipe sign-language tensorflow

Last synced: 7 days ago
JSON representation

Talking Hands is a project that uses MediaPipe and TensorFlow/Keras for real-time recognition of Brazilian Sign Language (LIBRAS). MediaPipe captures hand, facial, and body landmarks, which are processed by a TensorFlow-based neural network to transcribe signs into text and convert them into Portuguese audio.

• Host: GitHub
• URL: https://github.com/adrianocleao/talking-hands
• Owner: AdrianoCLeao
• License: mit
• Created: 2024-07-18T02:22:49.000Z (6 months ago)
• Default Branch: main
• Last Pushed: 2024-12-30T22:42:04.000Z (12 days ago)
• Last Synced: 2024-12-30T23:25:17.040Z (12 days ago)
• Topics: computer-vision, deep-learning, keras, libras, machine-learning, mediapipe, sign-language, tensorflow
• Language: Python
• Homepage:
• Size: 16.9 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# Brazilian Handsign Detection



This project aims to create a machine learning model that can identify, translate, and speak from live video of Libras (Brazilian Sign Language). It uses Google's Mediapipe to detect key points and Keras from TensorFlow to create the model.

## Project Structure

- `main.py`: Main file that initializes the GUI application for video recording and translation.

- `utils/`: Contains utility functions for data processing and keypoint detection.

  - `utils/utils.py`

  - `utils/constants.py`

  - `utils/text_to_speech.py`

- `models/`: Pre-trained models for gesture translation.

- `train/`: Scripts for model training.

  - `training_model.py`

  - `evaluate_model.py`

- `capture/`: Scripts for capturing and processing new data.

  - `capture_samples.py`

  - `create_keypoints.py`

## How to Use

1. **Create a Virtual Environment**

To isolate the project's dependencies, it is recommended to create a virtual environment using `venv`.

```bash

python -m venv venv

```

2. Activate the Virtual Environment

Activate the created virtual environment:

- On Windows:

```bash

venv\Scripts\activate

```

- On macOS and Linux:

```bash

source venv/bin/activate

```

3. Install the Requirements

With the virtual environment activated, install the project's requirements:

```bash

pip install -r requirements.txt

```

4. Run the Application

Execute the main.py file to start the GUI application:

```bash

python main.py

```

## Contributions

Feel free to contribute to this project. You can open issues and pull requests to discuss and implement improvements.

## License

This project is under the MIT license. Check LICENSE.md for more details.