Lithuanian Customs will make use of artificial intelligence to analyse goods being transported. A planned AI-based system will be able to evaluate X-ray images of cargoes and detect undeclared suspicious items.
Analysis by the system will be faster and more objective than analysis by humans. The goal is an analysis time of no more than 10 seconds per X-ray image. The system will handle some of the work that customs officials currently perform, leaving them more time for more complicated tasks. The plan is for the AI-based X-ray image analysis system to continue to evolve over time and for the algorithms to be trained with the information that is newly gathered.
“Our aim is to ensure effective border control: we believe that the advanced new system which is being developed will not only allow us to speed up the inspection of goods, but will also make inspections more objective,” says Darius Žvironas, the Director General of the Customs Department. He expects the innovations to yield more accurate analysis and so more transparent and efficient surveillance of national and international trade.
The new system will also make the cross-border transport of cargoes easier for honest businesses. It will speed up customs procedures and reduce the costs of submitting goods for customs control.
The project is being carried out by the software development companies of the Novian IT services group and Rubedo Systems. Launched in 2023, the project has a total value of EUR 2.231 million (including VAT), with completion planned in late 2025. It is part of a Customs Risk Management Modernisation project funded under the New Generation Lithuania Economic Recovery and Resilience Plan.
“Today AI is gaining ground in more and more areas, and it’s great that Lithuanian Customs is also making use of this advanced technology. We appreciate the innovativeness and importance of the project entrusted to us. For us, it is a broadening of the practical application of AI following projects we have already carried out in this area in the defence and aviation sectors,” says Tomas Vitkus, the CEO of the Novian group.
He said the project stands out not only for the high-reliability system that is being created and the use of a big data repository, but also because the algorithms that are being developed exploit the potential of graphics processing units (GPUs). The AI algorithms in development will help customs officials identify railway wagon types and numbers and detect undeclared suspicious items, and a tool developed by Novian during the project will allow faster tagging of data for training the AI.
Currently, the companies implementing the project have successfully completed the first phase, which involved analysing business processes as well as algorithms and data, selecting AI algorithms and designing the system architecture and interfaces. The project continues with programming work and the development of algorithms for detecting suspicious items.
