CNNs are a form of Neural Networks which are popular for the case of image and video classification. At a high-level CNNs apply an iterative pattern recognition method which scans individual images to extract key features that attempt to capture the latent properties which make up the image. CNNs can also be used for different time-series based/sequential analysis such as text classification or time-series forecasting.
These are the projects included in this repo utilizing CNNs. The projects are introductory explorations with commentary to explore different data prepatory and modeling techniques.
This implementation is sourced from a microsoft based Kaggle competition (https://www.kaggle.com/datasets/shaunthesheep/microsoft-catsvsdogs-dataset). The purpose of this project is for image classification of dogs vs cats. I explore the performance of an Artificial Neural Network vs the Convolutional Neural Network as well as identify potential steps in improving the performance of the models based on the images included in the input data.