Lunchlezing Atos

Analyzing x-ray scans of cargo trucks

We are Data Scientists from Atos. Within our work, the focus lies on analytics projects with the goal to obtain useful information and actionable insights out of raw data.

One of our projects revolves around the subject of image recognition. The purpose of this project is to research what possible methods can be implemented to apply image recognition to the customs check of cargo truck x-ray images. In this presentation we would like to take you with us in our process of extracting useful information out of raw data, in this case the cargo x-ray images.

We seek to answer questions like:

• How do we locate the container in an image?
• How full is the container?
• How many goods can we identify in the container?

We will show you the problems we run into during an image recognition project and the solutions we find in techniques such as feature extraction, edge detection, pattern recognition and texture segmentation.