Automated Explosive Detection Systems (EDS) utilizing Computed Tomography (CT) performs a series of X-ray scans
of the luggage being checked, then various 2D projection images of the luggage are generated from the collected data set
and sometimes 3D volumetric images of the luggage are generated in addition. Automatic explosives determination as to
the presence of an explosive in the luggage is determined through extensive data manipulation of the 2D and 3D image
sets, the results are then forwarded to a human interface for final review.
The final determination as to whether the luggage contains an explosive and needs to be searched manually is
performed by trained TSA (Transportation Security Administration) screeners following an approved TSA protocol. The
TSA protocol has the screeners visually inspect the projection images and the renderings of the automated explosive
results from detection to determine if the luggage needs to be suspected and consequently searched. Unlike conventional
X-ray systems, the user interface for EDS systems are usually designed to display one bag at a time. However, in airport
environments, there is usually more than one bag being processed. Therefore, segmentation is a crucial part of higher
quality screening. If the screeners have to manually manipulate (zoom, pan, separate) the image, this increases overall
screening time and decreases screener efficiency.
This paper presents a novel image segmentation technique that is geared towards, though not exclusive to, automated
explosive detection systems. The goal of this algorithm is to correctly separate each bag image to provide a higher
quality screening process while reducing the overall screening time and luggage search rates.