SINGLE-VIEW, MULTI-VIEW AND 3D IMAGING FOR BAGGAGE SCREENING: WHAT SHOULD BE CONSIDERED FOR EFFECTIVE TRAINING?

X-ray screening of passenger baggage is conducted at airports worldwide to prevent terrorist attacks and other acts of unlawful interference against civil aviation. Whether or not prohibited items are detected under X-ray examination depends both on the technology deployed and on human factors. While single-view imaging has been used for decades to screen passenger baggage, newer technology is based on multi-view and 3D imaging with automated explosives detection. However, the best equipment is of limited value if the humans who operate it are not trained appropriately[1]. Thus, what should be considered for training to be effective when single-view, multi-view or 3D imaging is used for passenger baggage screening? In this article, Adrian Schwaninger and Sarah Merks provide recommendations for cabin baggage and hold baggage screening based on scientific studies conducted with screeners over the last 15 years.

1. Cabin Baggage Screening

Cabin or carry-on baggage screening (CBS) refers to screening of bags and other belongings that passengers carry with them when going through an airport security checkpoint. The main prohibited item categories are guns, knives, improvised explosive devices (IEDs) and other prohibited items such as hand grenades[2,3]. Conventional CBS technology is based on single-view imaging. It displays cabin baggage from a single viewpoint and it does not feature automated explosives detection (Figure 1). The task of X-ray screeners is to visually inspect the X-ray images by searching for prohibited items and deciding whether the bag is harmless or not[4]. If screeners decide that a bag might contain a prohibited item, it is further inspected at the airport security checkpoint (secondary search) using explosives trace detection technology and physical search of the bag[5].

Newer CBS technology is based on multi-view imaging. It shows cabin baggage from two viewpoints, usually a main view and a side view (Figure 2). Such X-ray machines are also called explosive detection systems for cabin baggage screening (EDSCB). EDSCB highlight areas that might contain explosive material by coloured frames (Figure 2). The task of screeners is the same as when using single-view imaging, i.e. searching for prohibited items and deciding whether the bag is harmless or not[4]. In some countries, screeners conduct on-screen alarm resolution, i.e. they decide whether areas highlighted by the EDSCB are harmless or whether they might contain explosive material or an IED. In other countries, on-screen alarm resolution is not conducted and bags on which the EDSCB has alarmed are automatically sent to secondary search[7].

The newest CBS technology features 3D imaging that is based on computer tomography (CT). Such systems have better explosives detection (EDSCB) than multi-view imaging systems. With 3D imaging, cabin baggage is displayed as 3D rotatable images (Figure 3a). Most systems also provide cross-sectional images (also called slice or slab images, see Figure 3b for an example). Material that could be explosive is highlighted by colour (red areas in Figure 3a and Figure 3b). The task of X-ray screeners is the same as when using multi-view imaging.

1.1 General Training Recommendations for CBS

In CBS, screeners have to detect a large variety of items that could pose a threat during a flight. Recognising prohibited items in X-ray images can be difficult due to several reasons: 1) because screeners may not be familiar with certain objects based on their own life experience, 2) because many objects look very different in X-ray images, 3) because some prohibited items resemble everyday objects, and 4) because prohibited items can be difficult to recognise when depicted from unusual viewpoints[1]. Figure 4 illustrates these challenges.

Screeners must therefore not only learn which items are prohibited, but also what they look like in X-ray images in different rotations and how they can be distinguished from harmless objects[1,8]. Computer-based training (CBT) has been shown to be very effective and efficient in achieving and maintaining good detection performance of screeners in CBS[4,9,10]. A study conducted with 5,717 airport security screeners from more than 70 airports showed that CBT results in large increases of detection performance following a logarithmic function[10]. After 50 hours, CBT is still relevant, but mainly for maintaining detection performance. However, CBT systems do vary in effectiveness. For instance, Koller et al.[9] showed that an individually adaptive CBT that uses a large image library with target items in different rotations was very effective whereas a CBT that was not individually adaptive and had a smaller training library was not effective. Moreover, as terrorist threats evolve constantly, it is important to keep screeners up to date regarding new and emerging threats. Therefore, CBT libraries should be updated regularly based on systematic threat assessment[11].

“…a study conducted with 5,717 airport
security screeners from more than 70 airports showed that CBT results in large increases
of detection performance…”

1.2 Specific Training Recommendations for Multi-View CBS

As explained above, multi-view imaging for CBS displays cabin baggage from two different viewpoints (Figure 2). This can help screeners to detect prohibited items when depicted from unusual viewpoints or when superimposed by other items[6]. For example, the knife in Figure 2 is difficult to recognise in the main view (Figure 2a) whereas it is much easier to recognise in the side view (the knife is above the red rectangle in Figure 2b). Additionally, multi-view CBS highlights areas in the bag that could contain explosive material (EDSCB). A recently published study has shown that EDSCB is useful for increasing the detection of explosives and IEDs, in particular for screeners with limited experience [7]. However, in this study, screeners also showed a tendency to ignore many EDSCB alarms and they sometimes had difficulties in distinguishing explosives unconnected to other IED components from harmless organic mass. This challenge is illustrated in Figure 5.

In order to benefit from multi-view imaging for CBS, the following specific training recommendations can be made in addition to the general training recommendations mentioned above. Screeners should learn how the view of a bag from the side relates to the main view. Furthermore, screeners should be encouraged to use the side view, in particular if objects are depicted from difficult viewpoints and/or superimposed by other objects in the main view. If on-screen alarm resolution is used, screeners should receive training to increase their awareness for the fact that some explosives can be difficult to distinguish from harmless organic mass[7] and therefore secondary search using explosive trace detection and manual bag search is recommended using a risk-based approach when an EDSCB alarm cannot be resolved[5]. Finally, yet importantly, introducing new technology is an organisational change that can result in screeners reacting with concerns and resistance. Therefore, a phased approach allowing enough time for screeners to adapt to the new technology using a multi-view training simulator is recommended.

1.3 Specific Training Recommendations for 3D CBS

As explained above, 3D CBS technology allows 3D rotation of a bag (Figure 3a) and most systems also provide cross-sectional images (Figure 3b). These features may be even more effective than multi-view imaging at helping screeners to detect prohibited items when depicted from unusual viewpoints or when superimposed by other items. The main training challenge of introducing 3D imaging for CBS is that screeners need to learn how to effectively and efficiently use image display functions (3D rotation, cross-sectional images, etc.). It has not yet been investigated whether it is difficult for screeners to distinguish explosives from harmless organic masses using 3D CBS, as has been shown for 2D imaging for CBS[7]. However, based on observations at airports that have introduced 3D imaging, it seems that this challenge exists for 3D CBS as well.

In order to meet these challenges, the following specific training recommendations can be made for 3D imaging in addition to the general CBS training recommendations mentioned above. Screeners should learn how to use 3D rotation and other image display functions effectively and efficiently. If cross-sectional images are available, screeners should learn how they relate to the 3D rotatable view. Screeners should be encouraged to use 3D rotation, in particular if objects are depicted from difficult viewpoints and/or superimposed by other objects. If on-screen alarm resolution is used, screeners should receive training to increase their awareness that certain explosives are difficult to distinguish from harmless organic masses and therefore secondary search using explosive trace detection technology and manual bag search is recommended using a risk-based approach when an EDSCB alarm cannot be resolved[5]. As mentioned already for multi-view imaging, introducing new technology is an organisational change that can result in screeners reacting with concerns and resistance. This is particularly important when introducing 3D imaging for CBS as screeners have to learn new skills for operating the equipment (3D rotation of bags and using cross-sectional images if available). A phased approach allowing enough time for screeners to adapt to the new technology and learn the new skills needed for operating the equipment effectively and efficiently using a 3D simulator is highly recommended. This should consider differences between screeners’ abilities to operate computers and include enough supportive supervision.

“…screeners should learn what types of IEDs exist, how they look in X-ray images of hold baggage and how they can be distinguished from harmless objects…”

2 Hold Baggage Screening

Hold baggage screening (HBS) refers to screening of larger baggage that is stored in the hold of an aircraft during flights. Passengers have to register hold baggage at check-in stations before going through airport security checkpoints. Hold baggage is then processed by a baggage handling system containing X-ray machines that feature explosive detection systems (EDS). Although in CBS there are multiple target types (guns, knives, IEDs, explosives, other threats), this is not the case in HBS. Because passengers cannot access items stored in the hold of an aircraft, a gun or a knife does not pose a threat, and HBS targets only fully-functioning IEDs. Only X-ray images of hold baggage that has triggered an EDS alarm are sent to remote screening locations for on-screen alarm resolution by screeners. They visually inspect the X-ray images and decide whether the bag is harmless or whether it contains a fully-functioning IED with the following components: a triggering device, a power source, an explosive charge, and a detonator, all of which need to be connected to each other by, for example, wires (see Figure 6 for an example). If screeners decide that a piece of hold baggage might contain an IED, further baggage inspection is conducted using other HBS screening methods such as explosive trace detection, explosive detection dogs, passenger reconciliation, and the opening of bags[12]. While for CBS single-view imaging is still operational at many airports, multi-view or 3D imaging systems are used at most airports for HBS.

Multi-view imaging for HBS displays hold baggage from two viewpoints. Areas that might contain explosives are indicated by coloured frames (Figure 6). The task of HBS screeners is to resolve on-screen alarms, i.e. to visually inspect the X-ray images to decide whether the areas marked by the EDS contain a fully-functioning IED. It must be stressed that on-screen alarm resolution is the main task of HBS screeners. This is different to CBS, where on-screen alarm resolution is questionable[7]. For HBS, on-screen alarm resolution works very well if screeners receive appropriate training because the target items are fully functional IEDs only[12,13].

3D imaging for HBS is based on computer tomography (CT). It features 3D rotatable and cross-sectional images (also called slice or slab images, see Figure 7 for illustrations). The task of screeners is the same as with multi-view HBS: screeners conduct on-screen alarm resolution by visually inspecting the X-ray images to decide whether the areas marked by the EDS contain an IED. The main difference between 3D HBS and multi-view HBS is that in 3D HBS, screeners need to know how to use 3D rotation and how to display cross-sectional images. Some HBS systems combine 3D and multi-view imaging. Such systems show two 2D images with high resolution, a 3D rotatable image and a cross-sectional image on the computer monitor.

2.1 General Training Recommendations for HBS

IEDs can be made out of different types of explosives, triggering devices, power sources and detonators. Although fewer studies on the importance of training for HBS have been conducted, they have shown consistently that computer-based training (CBT) is very important in achieving and maintaining a high level of detection performance[14,15]. Therefore, general training recommendations for HBS are similar to those made for CBS: screeners should learn what types of IEDs exist, how they look in X-ray images of hold baggage and how they can be distinguished from harmless objects. HBS training should also be tailored to individual training needs by using an individually adaptive CBT[14,15]. Last but not least, the training library should be representative and up to date of current and emerging IED threats based on regular systematic threat assessment[11].

2.2 Specific Training Recommendations for Multi-View HBS

Because multi-view imaging for HBS displays hold baggage from two different view-points (Figure 6), this technology can help screeners to detect IEDs when depicted from unusual viewpoints or when superimposed by other items. This requires screeners to be able to use the side view in combination with the main view. Therefore, the following recommendations can be made in addition to the general training recommendations for multi-view HBS: screeners should learn how the side view relates to the main view. Screeners should also be encouraged to use the side view, particularly if objects are depicted from difficult viewpoints and/or superimposed by other objects in the main view.

2.3 Specific Training Recommendations for 3D HBS

3D imaging for HBS allows 3D rotation of a bag (Figure 7a) and movement through it with cross-sectional images (Figure 7b). These features are more effective than multi-view imaging at helping screeners detect IEDs when depicted from unusual viewpoints or when superimposed by other items. On the other hand, 3D imaging systems for HBS have lower image resolution than 2D multi-view imaging, which could result in lower detection performance. We have therefore conducted a study in which 2D and 3D screeners were tested with 2D multi-view and 3D imaging in a simulated hold baggage screening task[12]. The results are shown in Figure 8. 2D screeners achieved similar detection performance with both types of imaging. Features of 3D imaging systems (3D image rotation and cross-sectional images) seem to compensate for lower image quality. Visual inspection competency acquired with 2D imaging seems to transfer to visual inspection with 3D imaging. Interestingly, 3D screeners had higher detection performance than 2D screeners with both 2D and 3D imaging (Figure 8). A plausible explanation is that training and work experience with 3D imaging results in richer visual representations so that IEDs can be detected better with both types of imaging. In a follow-up study, we found that specific on-screen alarm resolution training could further increase detection performance of hold baggage screeners when using 3D imaging[13].

Similar to CBS, the main training challenge of introducing 3D imaging for HBS is that 3D rotation and using cross-sectional images requires screeners to learn new skills for operating the equipment. Screeners must learn how the 3D rotatable image and the cross-sectional image relate to each other. They should be encouraged to use these imaging functions in particular if objects are depicted from difficult viewpoints and/or superimposed by other objects.

---

Prof. Dr. Adrian Schwaninger is the head of the Institute Humans in Complex Systems (www.fhnw.ch/miks) of the School of Applied Psychology, University of Applied Sciences and Arts Northwestern Switzerland. He is also the chairman of the Center for Adaptive Security Research and Applications in Zurich (www.casra.ch).

---

Dr. Sarah Merks is senior research scientist and project manager at the Institute Humans in Complex Systems (www.fhnw.ch/miks) of the School of Applied Psychology, University of Applied Sciences and Arts Northwestern Switzerland. She also works at the Center for Adaptive Security Research and Applications in Zurich (www.casra.ch).

References

1. Schwaninger, A. (2005). Increasing efficiency in airport security screening. WIT Transactions on the Built Environment, 82, 407–416.
2. Schwaninger, A., Hardmeier, D., & Hofer, F. (2005). Aviation security screeners visual abilities & visual knowledge measurement. IEEE Aerospace and Electronic Systems, 20(6), 29-35.
3. Koller, S., & Schwaninger, A. (2006). Assessing X-ray image interpretation competency of airport security screeners. Proceedings of the 2nd International Conference on Research in Air Trans-portation, ICRAT 2006, Belgrade, Serbia and Montenegro, June 24-28, 2006, 399-402.
4. Koller, S., Drury, C., & Schwaninger, A. (2009). Change of search time and non-search time in X-ray baggage screening due to training. Ergonomics, 52(6), 644–656.
5. Sterchi, Y., & Schwaninger, A. (2015). A first simulation on optimizing EDS for cabin baggage screening regarding throughput. Proceedings of the 49th IEEE International Carnahan Confer-ence on Security Technology, Taipei, Taiwan, September 21-24, 2015, 55-60.
6. von Bastian, C. C., Schwaninger, A., & Michel, S. (2008). Do multi-view X-ray systems improve X-ray image interpretation in airport security screening? Zeitschrift für Arbeitswissenschaft, 3, 166–173.
7. Hättenschwiler, N., Sterchi, S., Mendes, M., Schwaninger, A. (2018). Automation in airport securi-ty X-ray screening of cabin baggage: Examining benefits and possible implementations of auto-mated explosives detection. Applied Ergonomics, 72, 58-68.
8. Schwaninger, A. (2004). Computer based training: the enhancement of human factors. Aviation Security International, 2004(2), 31-36.
9. Koller, S., Hardmeier, D., Michel, S., & Schwaninger, A. (2008). Investigating training, transfer and viewpoint effects resulting from recurrent CBT of x-ray image interpretation. Journal of Transportation Security, 1(2), 81–106.
10. Halbherr, T., Schwaninger, A., Budgell, G. R., & Wales, A. W. J. (2013). Airport security screener competency: A cross-sectional and longitudinal analysis. International Journal of Aviation Psychology, 23(2), 113–129.
11. Schmid, F. (2015). Systematic Threat Assessment: how to adapt security measures based on the research on latest threats. International Airport Review, 19(4), 60-63.
12. Hättenschwiler, N., Mendes, M., & Schwaninger, A. (2018). Detecting bombs in X-ray images of hold baggage: 2D versus 3D imaging. Human Factors, 61(2), 305-321.
13. Hättenschwiler, H., Merks, S., & Schwaninger, A. (2018). Airport security X-ray screening of hold baggage: 2D versus 3D imaging and evaluation of an on-screen alarm resolution protocol. Pro-ceedings of the 52nd IEEE Carnahan Conference on Security Technology, Montreal, Canada, 22-25 October, 2018, 1-5.
14. Schwaninger, A., & Hofer, F. (2004). Evaluation of CBT for increasing threat detection perfor-mance in X-ray screening. In K. Morgan & M. J. Spector (Eds.), The Internet society: Advances in learning, commerce and security (pp. 147–156). Southampton, England: WIT Press.
15. Schwaninger, A., Hofer, F., & Wetter, O. E. (2007). Adaptive computer-based training increases on the job performance of x-ray screeners. Proceedings of the 41st Carnahan Conference on Security Technology, Ottawa, Canada, October 8–11, 2007, 117-124.