Understanding convolutional neural networks using a minimal model for handwritten digit recognition

The contribution of this paper is to bridge the gap on understanding the mathematical structure and the computational implementation of a convolutional neural network (CNN) using a minimal model (Minimal CNN). The proposed minimal CNN is presented using a layering approach. This approach provides a concise and accessible understanding of the main mathematical operations of a CNN. Hence, it benefits beginners and non-mathematical prolific learners to acquire foundational knowledge that informs a principled understanding of CNNs without having an intimidating experience. A handwritten digit recognition using MNIST handwritten digit dataset is used to experiment the performance of the proposed minimal CNN with other neural networks.