At present, there is a growing abundance of weather and climate model output as well as satellite data both in time and dimensionality. This provides greater details into cyclones that have not been observed before but also creates a problem of quickly utilizing these data in real time due to their size and frequency. To tackle this problem, we developed a method for extratropical and tropical cyclone image recognition using deep learning (DL) on Global System Forecasting (GFS) analysis weather model output as well as on Geostationary Operational Environmental Satellite (GOES) satellite imagery. DL is a type of machine learning that initially runs with large quantities of well-labeled data to train itself and then can be used without the initial training process on unlabeled datasets to very quickly identify regions of interest (ROI). There are numerous ways to design the architect of a DL model and for our purposes, we selected the UNET because of its previous success in image tasks. In order to train our UNET DL model, we engineered a two-step combination of a heuristic-based model and a DL model. We designed a heuristic model to derive the labeled data required for training in supervised DL models. Once trained, the DL model then runs quickly on unlabeled data to extract regions of cyclone activity.