Manipulating a Tetris-Inspired 3D Video Representation

TL;DR

This paper explores various 3D data representations and packing algorithms for video synopsis, aiming to improve video compression techniques in surveillance by leveraging a Tetris-inspired approach.

Contribution

It introduces a formal framework for 3D video synopsis, discusses modifications for simplified problems, and applies packing algorithms inspired by Tetris to optimize video compression.

Findings

Effective 3D packing algorithms for video synopsis

Comparison of different spatio-temporal data representations

Discussion of future research directions in 3D video compression

Abstract

Video Synopsis is a technique that performs video compression in a way that preserves the activity in the video. This technique is particularly useful in surveillance and monitoring applications. Although it is still a nascent field of research, there have been several approaches proposed over the last two decades varying with the application, optimization type, nature of data feed, etc. The primary data required for these algorithms arises from some sort of object tracking method. In this paper, we discuss different spatio-temporal data representations suitable for different applications. We also present a formal definition for the video synopsis algorithm. We further discuss the assumptions and modifications to this definition required for a simpler version of the problem. We explore the application of a packing algorithm to solve the problem of video synopsis. Since the nature of the data is three dimensional, we consider 3D packing problems in the discussion. This paper also provides an extensive literature review of different video synopsis methods and packing problems. Lastly, we examine the different applications of this algorithm and how the different data representations discussed earlier can make the problem simpler. We also discuss the future directions of research that can be explored following this discussion.