Bornomala

An experimental app for web that performs real-time bangla vowel handwritten character recognition captured using web canvas.

Project Demo Video

Technologies

• Deep Learning

  • Convolution Neural Network (CNN)
  • Python, keras, tensorflow, opencv, numpy

• Web

  • Flask (python web framework)
  • Jquery, Ajax, Bootstrap

Dataset

A data file is required for every language you want to recognize. The dataset was obtained online from the CMATERdb pattern recognition database repository. It consists of a Train folder and a Test folder, containing 12,000 and 3,000 images respectively. We only used vowels in our work.

• For training data, we found 2112 images belonging to 11 classes.
• For test data, we found 528 images belonging to 11 classes.

CNN Model Summary

 _________________________________________________________________
 Layer (type)                 Output Shape              Param #   
 =================================================================
 conv2d_1 (Conv2D)            (None, 30, 30, 32)        320       
 _________________________________________________________________
 max_pooling2d_1 (MaxPooling2 (None, 15, 15, 32)        0         
 _________________________________________________________________
 conv2d_2 (Conv2D)            (None, 13, 13, 64)        18496     
 _________________________________________________________________
 batch_normalization_1 (Batch (None, 13, 13, 64)        256       
 _________________________________________________________________
 max_pooling2d_2 (MaxPooling2 (None, 6, 6, 64)          0         
 _________________________________________________________________
 dropout_1 (Dropout)          (None, 6, 6, 64)          0         
 _________________________________________________________________
 conv2d_3 (Conv2D)            (None, 4, 4, 128)         73856     
 _________________________________________________________________
 batch_normalization_2 (Batch (None, 4, 4, 128)         512       
 _________________________________________________________________
 max_pooling2d_3 (MaxPooling2 (None, 2, 2, 128)         0         
 _________________________________________________________________
 dropout_2 (Dropout)          (None, 2, 2, 128)         0         
 _________________________________________________________________
 flatten_1 (Flatten)          (None, 512)               0         
 _________________________________________________________________
 dense_1 (Dense)              (None, 256)               131328    
 _________________________________________________________________
 dense_2 (Dense)              (None, 512)               131584    
 _________________________________________________________________
 dense_3 (Dense)              (None, 100)               51300     
 _________________________________________________________________
 dense_4 (Dense)              (None, 11)                1111      
 =================================================================
 Total params: 408,763
 Trainable params: 408,379
 Non-trainable params: 384
 _________________________________________________________________

Installation

To build and run the app, first of all clone this project.

You may also consider installing the following dependencies:

• TensorFlow.
• Keras.
• Open CV.
• Flask.

Then open cmd in the project directory and run the command python main.py.