Skip to Main Content
This paper presents a multi tier web architecture that integrates web technology, Database system, and Batch processing tools for the development of a real time threat detection system. Four data repository models are introduced for effective data storage and retrieval. The baseline feature vectors are introduced and stored in the database table using a batch job. The batch job performs load balancing by calculating the new feature vector using the offline server and updates the online database server. The illustrative application uses the Hierarchical Multi level HVS segmentation, ratio based edge detection, and support vector machine for threat recognition and detection. The 64 bit edge based feature vector is generated for the baseline images and the input test object images using the cell edge distribution approach. The experimental results demonstrate that the presented framework is efficient in facilitating accurate threat detection and support the development of portable, reusable and scalable object recognition applications for heterogeneous distributed environment.