Temporal Unknown Incremental Clustering (TUIC) Model for Analysis of Traffic Surveillance Videos

TL;DR

The paper introduces TUIC, a real-time, non-parametric clustering model using Gibbs sampling for traffic video analysis, effectively tracking moving objects across frames for abnormality detection.

Contribution

Proposes TUIC, a novel Gibbs sampling-based incremental clustering method for real-time pixel motion analysis in traffic videos, enhancing scene representation and abnormality detection.

Findings

Fast clustering with Θ(kn) complexity

Accurate motion clustering in traffic videos

Potential for real-time traffic analysis

Abstract

Optimized scene representation is an important characteristic of a framework for detecting abnormalities on live videos. One of the challenges for detecting abnormalities in live videos is real-time detection of objects in a non-parametric way. Another challenge is to efficiently represent the state of objects temporally across frames. In this paper, a Gibbs sampling based heuristic model referred to as Temporal Unknown Incremental Clustering (TUIC) has been proposed to cluster pixels with motion. Pixel motion is first detected using optical flow and a Bayesian algorithm has been applied to associate pixels belonging to similar cluster in subsequent frames. The algorithm is fast and produces accurate results in $\Theta(kn)$ time, where $k$ is the number of clusters and $n$ the number of pixels. Our experimental validation with publicly available datasets reveals that the proposed framework has good potential to open-up new opportunities for real-time traffic analysis.