Mina Kheirabadi: CL2018 is a project funded through Innovation fund Denmark in order to improve the security x-ray screening with using some image analysis and machine learning techniques. The project will be involved in different parts.
In the first step, we can have 3D image reconstructed by images taken from different angles. DTU has already the reconstruction algorithm. In the second part, we would like to improve threat detection by using many projections (Multi-views) as well as Multi-energy detectors. Using many projections can provide us with the advantages of seeing the objects from different views especially when they are cluttered.
In this project, Multiple energy-selective measurements of the attenuation are taken that compared to dual energy detectors can give us much more information in order to detect threat items with high accuracy. In addition, we would like to decrease the number of projections to 4 by using some segmentation information. Therefore, this project will be involved in some methods in segmentation, classification and object detection to improve security application.
PhD projekt: 2016 -
Supervisor: Anders B. Dahl, Section for Visual Computing