Original article was published by BaggageAI on Artificial Intelligence on Medium
Disrupting Conventional Baggage Scanning with AI
Baggage Scanning with AI is the best solution to overcome the traditional method of manual baggage scanning.
Defining the Problem
In spite of the consistent increase in air, rail and road traffic over the years, the technology supporting the security of these facilities, has not grown simultaneously. This has led to checkpoints that are often clogged because of the conventional and manual baggage scanning techniques.
Moreover, there is a lot of scope for human errors to crawl into the system. As can be easily seen in any major checkpoint, X-ray baggage machines are almost always accompanied by human screeners who pass the final judgement after inspecting the scans on screens. This process is effective, but only for a short while and only if the number of people passing through the point is very less.
Otherwise, long hours of screen exposure can exhaust any individual, leading to a cloudy judgement. Moreover, time is a giant factor (baggage screeners spend almost 10 hours every day speculating the screens). Human screening definitely eats up a lot of hours and causes delays which are sometimes financially disruptive.
Previous Approach Towards Solution Development
Over the years, like in many parallel technologies, the security techniques have also grown in an incremental fashion. What this means is, the core technology of the baggage scanner remains constant and only a few surface-level adjustments are made in the equipment over time. This serves the purpose but is definitely not bringing a significant or visible change.
How Can Innovation Help?
In order to revolutionize the system such that the giant gap between real traffic and the backdated technology can be met, is to bring in Artificial Intelligence. AI is doing wonders across all sectors, then why restrict it from reforming our conventional X-ray baggage inspection systems as well?
With advanced Baggage Scanning with AI, the jump from incremental to exponential growth is highly viable now. BaggageAI’s scanner add-on offers a dual solution to get rid of the two most recurring problems in Baggage Scanner Machines — low throughput and low accuracy in the detection of threats.
The first problem of low throughput is solved with the time-optimized technology of baggage scanning in contrast to conventional technology.
The second problem of low accuracy is solved by using the precise AI detection system which is trained on a large variety of data with optimum computer vision technology. As observed in a pilot project with Airport Authority of India in Pune, the BaggageAI solution was 98% accurate (threats and non-threats correctly detected) and 94% precise (threats correctly detected) when used with the airport baggage x-ray machines.
The major misconception against not just AI Scanning technology, but against AI technology in general, is that it will replace the human workers and leave them with no source of income to fend for their families.
If we trace back the origins of AI, it can be found that AI was brought to the table, not to replace the diners, but only to assist them in enjoying a meal better.
BaggageAI has stuck to this core ideology of Artificial Intelligence to create a system that aids and stimulates human works to perform better and more efficiently. It is definitely not a replacement since it will be reckless to think that critical situations which deal with the lives of millions of people can be placed solely in the hands of a few machines.
BaggageAI instead, eliminates the scope of errors which result due to exhaustion and fatigue. It is a replacement only for the shortcomings of humans, but definitely not a replacement for the strength of human screeners.
When we saw that people were prone to common errors like carrying liquids or prohibited gadgets or missing threats when it came to screeners, most often unintentionally, we realized that the accumulation of such small errors can lead to precarious events endangering a lot of innocent people.
Therefore, we came up with the idea of eliminating these small errors, one at a time.
However, there were also more critical things to look after — for instance when more dangerous objects were smuggled through careful planning of organized groups through loosely managed baggage scanner machines.
But overall, it boiled down to the same solution. If we could succeed in detecting the minor threats, we could build a system based on the same structure that could identify the major threats. That was how BaggageAI was born. A robust add-on to pump up security!
A stitch in time saves nine. Literally.