Computer system has learned to recognize x-ray images and alerts if it “sees” illegal items.

BitRefine Group, the leading company in machine learning and computer vision, introduces solution for the classification of x-ray baggage images using state of the art convolutional neural networks. The computer vision system analyzes x-ray images by their shapes and automatically gives alert when it “sees” an illegal item. This leads to the reduction of labor-intensive screening. This, by far, is the greatest evidence that with continuous research and development, fully autonomous x-ray luggage screening is possible in the future.

X-ray baggage security screening is widely used to maintain aviation and transport security, itself posing a significant image-based screening task for human operators reviewing compact, cluttered and highly varying baggage contents within limited time-scales. Within both increased passenger throughput in the global travel network and an increasing focus on wider aspects of extended border security (e.g. freight, shipping postal), this posed both a challenging and timely automated image classification task.

Generally, x-ray scanner operators work 40 hours a week and ideally, they should not continuously monitor images for periods longer than 20 minutes in any hour (In reality, operators may have to do the job on longer periods). Given these facts, there is a probability that there would be lapses in detecting illegal items inside the baggage because keeping focus on the screen for 20 minutes per hour is a challenging work. 

X-ray AirportThe system eliminates the need of tedious and time-consuming manual baggage inspection. This doesn’t only speeds up the process but also serves as a security assurance to passengers. This object-detection system can work with objects such as: firearms, stun guns, sprays, knives and scissors, screwdrivers, batteries, bottles and more!

Deep neural networks for object detection

The task of automated threat screening in x-ray baggage imagery has always been a challenge in terms of classification and detection of items. The modernized version of multi-layer perceptrons is the deep convolutional neural networks (CNN). It has been used in various ways such as speech recognition and natural language processing. More importantly, it’s used in the field of computer vision for tasks such as image classification, object detection and segmentation. With the availability of large data sets, solution for the classification of x-ray baggage images using state of the art convolutional neural networks is made possible. Based on research, CNN achieved superior performance achieving 99% True Positive.

Future development will be focused on accumulating larger datasets containing various baggage objects will lead to fully automated scheme for a real time application. Plus development will be focused on the classification of additional object types, the use of different X-ray imagery (capturing more images per item) and an investigation into further optimization technique. Machine learning and computer vision technology develop rapidly. Specialists from BitRefine believe that in a decade all screening devices will work either fully or semi-autonomously. 

About BitRefine

Bitrefine group solves clients’ toughest challenges in the field of machine learning and computer vision. We offer our clients solutions that will help them stay ahead of trends and maintain a competitive advantage in the market. Deep data analysis, visual information comprehension, robotics, automation, medical image processing, behavior prediction – this is the short list of application that we develop. We partner with clients to turn emerging technology into a real-world product.