Utilized an LSTM over a CNN to create a model for live video classification. The CNN used was a retrained version of Google's Inception v3 for image classification. The model was retrained using 500+ hours of footage, and deployed on a multithreaded web platform for system usage and management.