Learning Semantic Scene Models by Trajectory Analysis

DMCA

Learning Semantic Scene Models by Trajectory Analysis (2006)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Xiaogang Wang , Kinh Tieu , Eric Grimson

Venue:	In ECCV (3) (2006
Citations:	90 - 11 self

BibTeX

@INPROCEEDINGS{Wang06learningsemantic,
    author = {Xiaogang Wang and Kinh Tieu and Eric Grimson},
    title = {Learning Semantic Scene Models by Trajectory Analysis},
    booktitle = {In ECCV (3) (2006},
    year = {2006},
    pages = {110--123}
}

OpenURL

Abstract

Abstract. In this paper, we describe an unsupervised learning framework to segment a scene into semantic regions and to build semantic scene models from longterm observations of moving objects in the scene. First, we introduce two novel similarity measures for comparing trajectories in far-field visual surveillance. The measures simultaneously compare the spatial distribution of trajectories and other attributes, such as velocity and object size, along the trajectories. They also provide a comparison confidence measure which indicates how well the measured image-based similarity approximates true physical similarity. We also introduce novel clustering algorithms which use both similarity and comparison confidence. Based on the proposed similarity measures and clustering methods, a framework to learn semantic scene models by trajectory analysis is developed. Trajectories are first clustered into vehicles and pedestrians, and then further grouped based on spatial and velocity distributions. Different trajectory clusters represent different activities. The geometric and statistical models of structures in the scene, such as roads, walk paths, sources and sinks, are automatically learned from the trajectory clusters. Abnormal activities are detected using the semantic scene models. The system is robust to low-level tracking errors. 1

Keyphrases

semantic scene model trajectory analysis true physical similarity unsupervised learning framework novel clustering algorithm abnormal activity novel similarity measure comparison confidence different activity far-field visual surveillance semantic region image-based similarity spatial distribution trajectory cluster longterm observation comparison confidence measure statistical model velocity distribution similarity measure object size low-level tracking error different trajectory cluster

DMCA