To implement handwritten digit recognition using neural networks. This research work analyses the behaviour of classification techniques (CNN) in a large handwriting dataset (MNIST) to predict a digit.
https://www.kaggle.com/datasets/hojjatk/mnist-dataset
Convolutional Neural Networks are categorized as deep artificial neural networks. It isused for image recognition and also in object recognition. It has been used under various other applications in detection algorithms. CNN’s core building block is the convolutional layer. This layer parameter is composed of kernels (also known as learnable filters) which have a not-so-large receptive field but extend through the full depth of the input volume. When the forward pass is applied, each filter is convolved across the width and height of the input and then computing the dot product and then developing a 2-D activation map of the corresponding filter. As a result, the network learns when they see certain types of features at a spatial location in the input. Activation maps are then given in the lower sample layer and, as a decision, this method is applied to a patch time. CNN has a fully connected layer, which classifies the output with a label per node.
