Fast methods for frame prediction

Thesis Proposal Details

Supervisor: Marco Cagnazzo

Creation Date: 09/10/2025 11:43

Description

In this thesis, we explore state-of-the-art methods for future frame prediction. 

After the bibliographical research, the goal is to set up the most recent methods and compare them, considering not only the image quality but also the execution time in real systems.

Dataset and methods

Dataset type: Already acquired data

Dataset description: We use public video databases

List of Methods: Neural networks Video analysis Motion estimation

Preparatory Courses

Multimedia Communications Neural networks and deep learning

Tags
deep learning neural networks prediction video
Back to proposals list