SpeechEase

Empowering Clear Communication, One Word at a Time

Comment

Inspiration

I happened to watch a bollywood movie 'black' wherein a teacher helps a student to learn words through her disabilities of being blind and deaf, that movie was poignant. In the real world it's difficult to find teachers like those and such children don't get help. this project is my effort in this direction to help these special children that deserve to learn just like we do.

What it does

SpeechEase can offer several functionalities to assist individuals with speech difficulties. Here are some key features:

Practice Speaking:

Users can engage in guided speaking exercises tailored to their specific speech challenges. This feature may include prompts and feedback to help improve clarity and confidence in their speech.

Learn Pronunciation:

The application can provide phonetic guidance and audio examples to help users learn the correct pronunciation of words. Users can practice alongside these examples and receive feedback on their pronunciation.

Word Difficulty Notebook:

Users can maintain a personalized library of words or phrases they find challenging to pronounce. This notebook can be a useful tool for focused practice and tracking progress over time.

Progress Tracking and Feedback:

SpeechEase can track users’ progress in their speaking practice and provide feedback on areas of improvement. This feature could include metrics on pronunciation accuracy, fluency, and areas needing more practice.

How we built it

Frontend: React.js: For building user interfaces, allowing for dynamic and responsive web applications. Tailwind CSS: For styling, enabling a clean and modern UI with utility-first CSS.

Backend: Flask: A lightweight Python web framework to handle API requests and serve the application. Machine Learning Libraries: Such as scikit-learn, librosa, or TensorFlow for speech analysis and model training.

Challenges we ran into

the main challenge that i ran into training the ml model to identify the voice and find the issuses in a short period of time , Even though the feature is not fully functional still i am proud that i trained the model on the huggface dataset. also the vercel only has the fornt end part of the website where there is problem loading the svgs.

Built With

Share this project:

Updates