Video Frame Prediction using Deep Learning

Abstract

Advances in Deep Learning helped in developing exiting applications in the past few years like Style Transfer, Image Generation. It has helped us achieving super human performance and state-of-the-art results in various tasks like Object Recognition/ Classi cation. It has established itself as a go-to technique for solving variety of tasks like Machine Translation in Natural Language Processing to Image Classi - cation, Image Captioning in Computer Vision etc. Future Frame prediction is one of the problems in Computer Vision that received a lot of interest in the recent past for its use in Autonomous Driving, Weather Forecasting, Tra c Estimation etc. Next Frame Prediction is the focus of this thesis. This work will explain the fundamentals of future frame prediction and it will give an overview of existing approaches and present an approach to solve the problem.