Invariant Feature Extraction and Biased Statistical Inference for Video Surveillance

Using cameras for detecting hazardous or suspicious events has spurred new research for security concerns. To make such detection reliable, researchers must overcome difficulties such as variation in camera capabilities, environmental factors, imbalances of positive and negative training data, and asymmetric costs of misclassifying events of different classes. Following up on the event-detection framework that we proposed in [12], we present in this paper the framework’s two major components: invariant feature extraction and biased statistical inference. We report results of our experiments using the framework for detecting suspicious motion events in a parking lot.

• Type: Conference
• Year: 2003
• Where: AVSS
• Authors: Yi-Leh Wu, Long Jiao, Gang Wu, Edward Y. Chang, Yuan-Fang Wang
• Keywords: AVSS 2003, Invariant Feature Extraction, Signal Processing, Suspicious Events, Suspicious Motion Events