Handwritten Digits Recognition (CNN)

You can upload your own image and see what the model says about it.

We can improve this model by changing the model architecture, and adding more training data. At this level I wanted to showcase my work towards implementation and deployment of Deep Learning model.

Development summary

High-level architecture of the application

Model summary

Training accuracy: 100% Validation accuracy: 99% 
_________________________________________________________________
Layer (type)                Output Shape              Param #   
=================================================================
conv2d (Conv2D)             (None, 26, 26, 32)        320       
                                                                
max_pooling2d (MaxPooling2  (None, 13, 13, 32)        0         
D)                                                              
                                                                
conv2d_1 (Conv2D)           (None, 11, 11, 64)        18496     
                                                                
dropout (Dropout)           (None, 11, 11, 64)        0         
                                                                
conv2d_2 (Conv2D)           (None, 9, 9, 64)          36928     
                                                                
max_pooling2d_1 (MaxPoolin  (None, 4, 4, 64)          0         
g2D)                                                            
                                                                
flatten (Flatten)           (None, 1024)              0         
                                                                
dense (Dense)               (None, 100)               102500    
                                                                
dense_1 (Dense)             (None, 10)                1010      
                                                                
=================================================================
Total params: 159254 (622.09 KB)
Trainable params: 159254 (622.09 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________