Handwritten Bengali Vowel Recognition Web App

🔗 Deployed Link: Handwritten Bengali Vowel Recognition Web App

Tech Stack

• Python
• TensorFlow
• Gradio
• NumPy

Features

Input Method

• Interactive sketchpad captures handwritten input in real-time.

Model

• CNN trained to classify 11 Bengali vowel characters.
• Trained with TensorFlow on image dataset of handwritten characters.

Interface

• Live sketchpad input with prediction shown instantly.
• Clean UI built using Gradio.

Workflow

1. Data Collection
   Dataset of handwritten Bengali vowels prepared and augmented.

2. Model Training
   CNN trained with categorical cross-entropy and accuracy metrics.

3. Web App Deployment
   Gradio app created for real-time predictions from user-drawn input.

4. User Interaction
   Users draw a vowel, and the app instantly returns the predicted character.