Single Energy X-ray Image Colorization Using Convolutional Neural Network for Material Discrimination

Colorization in X-ray material discrimination is considered one of the main phases baggage inspection systems for detecting contraband and hazardous materials by displaying different with specific colors. The substructure identifies based on their atomic number. However, images are checked assigned a human factor, which may decelerate verification process. Therefore, researchers used computer vision machine learning methods to expedite examination process ascertain precise identification elements. This study proposes color-based method single-energy dual-energy colorization. We use convolutional neural network discriminate into several classes, such as organic, non-organic substances, metals. It highlights details objects, including occluded compared commonly segmentation methods, do not show objects. trained tested our model three popular datasets, Korean datasets comprising kinds scanners: (Rapiscan, Smith, Astrophysics), SIXray, COMPASS-XP. results showed that proposed achieved high performance colorization terms peak-signal-to-noise ratio (PSNR), structural similarity index (SSIM), learned perceptual image patch (LPIPS). applied models we obtained from each model.