Admin
Joined - December 2025
Key Features:
Drawing Canvas:
Real-time drawing with adjustable brush size
Smooth line drawing
Clear canvas functionality
Model Management:
Multiple model architectures (CNN, Dense, Custom CNN)
Train on MNIST dataset
Save/load trained models
Real-time training progress
Recognition Features:
Real-time character prediction
Confidence scores
Visual confidence bars for top predictions
Preprocessing pipeline (resize, invert, normalize)
Advanced GUI Components:
Paned windows for resizable panels
Status bar with updates
Training history visualization
File dialog for saving/loading
Image Handling:
Save drawn images
Load external images for recognition
Automatic preprocessing
Installation Requirements:
bash
pip install tensorflow scikit-learn pillow matplotlib numpy To Use:
Draw a digit (0-9) on the canvas
Click Recognize to see the prediction
Train your own model with different architectures
Save/Load models for future use
View training history graphs
Model Options:
CNN - Basic convolutional neural network
Dense - Fully connected neural network
Custom CNN - Deeper convolutional network
Get the updates.
Up to Top
Comments