R4-B.1: Toward Advanced Baggage Screening: Reconstruction and Automatic Target Recognition (ATR)

In transportation security applications, Computed Tomography (CT) scanners are widely used to scan checked baggage for threatening materials. Traditionally, images are reconstructed using direct methods, such as fi ltered back projection (FBP). Model-based iterative reconstruction (MBIR) potentially offers many important advantages over traditional methods for the security screening of checked baggage. It has the potential to reduce metal artifacts and improve resolution. All these improvements have the potential to improve the detection/false alarm tradeoff for CT security screening systems. Furthermore, automatic target detection and recognition from the scanned images can reduce the cost of human labors, help to extract important information and support human judgments. The objective of this research is to investigate the potential of MBIR algorithms for the security screening application. In addition, we aim to develop a new Automatic Target Recognition (ATR) system which will incorporate advanced segmentation, feature extraction and classifi cation techniques in order to improve the current state-of-the-art ATR performance. To do this, we improved the existing MBIR algorithms, which are based on a monochromatic model for X-ray data, by estimating a polynomial for correction of nonlinear effects simultaneously with the image. We also built our new ATR system, which advances segmentation and classifi cation over the standard software provided by ALERT. During the last project period, we successfully achieved better reconstruction images, which reduce artifacts caused by nonlinear attenuation behavior, such as beam hardening, scatter and other incompletely modeled attributes of the data and improved detection/false alarm score in the ATR system. We will continue to develop new techniques in order to further improve performance for challenging cases, such as images with cluttered objects and metal streaking artifacts.