With the use of machine & deep learning algorithms, tasks ranging from object recognition in images to adding sound to silent films may now be completed with greater ease than ever before. Similar to this, Recognition of handwritten text is a key field for advancement and research with many different possible outcomes. The capability of a computer to accept and interpret understandable handwritten input from sources such as pictures, touch-screens, paper documents, etc. is known as handwriting recognition (HWR). Evidently, utilising MNIST datasets, artificial neural networks (MLP), convolution neural networks (CNN) we conducted handwritten digit recognition in this research. To find the most effective model for digit recognition, our major goal is to compare the training and validation accuracy and loss of the models mentioned above.
Key words: Machine Learning, Deep Learning, Handwritten Digit Recognition, Convolution Neural Network (CNN), Artificial Neural Network (Multi-Layer Perceptron), and MNIST datasets.
|
