Article

Permanent URI for this collectionhttps://hdl.handle.net/10361/7030

Browse

Now showing 1 - 3 of 3

Automated intruder detection from image sequences using minimum volume sets
(© 2012 International Journal of Communication Networks and Information Security, 2012) Ahmed, Tarem; Wei, Xianglin; Ahmed, Supriyo Sabbir; Pathan, Al-Sakib Khan; Department of Electrical and Electronic Engineering
We propose a new algorithm based on machine learning techniques for automatic intruder detection in visual surveillance networks. The proposed algorithm is theoretically founded on the concept of Minimum Volume Sets. Through application to image sequences from two different scenarios and comparison with existing algorithms, we show that it is possible for our proposed algorithm to easily obtain high detection accuracy with low false alarm rates.
Efficient and effective automated surveillance agents using kernel tricks
(© 2012 The Society for Modeling and Simulation International., 2012) Ahmed, Tarem; Pathan, Al-Sakib Khan; Ahmed, Supriyo Sabbir; Wei, Xianglin; Department of Electrical & Electronic Engineering
Many schemes have been presented over the years to develop automated visual surveillance systems. However, these schemes typically need custom equipment, or involve significant complexity and storage requirements. In this paper we present three software-based agents built using kernel machines to perform automated, real-time intruder detection in surveillance systems. Kernel machines provide a powerful data mining technique that may be used for pattern matching in the presence of complex data. They work by first mapping the raw input data onto a (often much) higher-dimensional feature space, and then clustering in the feature space instead. The reasoning is that mapping onto the (higher-dimensional) feature space enables the comparison of additional, higher-order correlations in determining patterns between the raw data points. The agents proposed here have been built using algorithms that are adaptive, portable, do not require any expensive or sophisticated components, and are lightweight and efficient having run times of the order of hundredths of a second. Through application to real image streams from a simple, run-of-the-mill closed-circuit television surveillance system, and direct quantitative performance comparison with some existing schemes, we show that it is possible to easily obtain high detection accuracy with low computational and storage complexities.
Automated surveillance in distributed, visual networks: an empirical comparison of recent algorithms
(© 2014 Science and Engineering Research Support Society, 2014) Ahmed, Tarem; Ahmed, Supriyo Sabbir; Pathan, Al-Sakib Khan; Department of Electrical and Electronic Engineering
A number of algorithms have been recently proposed for automatic intruder detection from CCTV images. Past researchers have typically tested these algorithms on centralized networks where all images are transmitted to a central control room. This paper demon- strates the applicability of a selection of such algorithms to a distributed network of wireless sensors. A distributed network of wireless visual sensors was simulated using a number of high-resolution webcams setup in the hallways of an academic building. The selected algo- rithms were then applied in a distributed fashion at each node. An empirical comparison of the most popular of the recent algorithms on a simulation of a wireless sensor network was thus obtained. This paper provides corroborating evidence in support of the most effective of such algorithms to the problem of automatic anomaly detection from image streams.

Browse

Recent Submissions