Enhanced threat object detection through generative adversarial networks based synthetic X-ray image generation

2026-06-12

PRE

Archana Singh *

Michael Lewis

null Dhiraj

DOI：10.1016/j.imavis.2026.106071

Original

Original request for help

Save

Abstract

En 中文

• PCA-based orientation grouping reduces pose variation in threat images. • Modified WGAN-GP with residual blocks, self-attention, and FM loss for realistic samples. • Attention-guided discriminator improves threat image generation quality. • Novel sample selection identifies the most realistic generated threats. • Density-based threat insertion creates realistic augmented X-ray datasets.

Enhanced threat object detection through generative adversarial networks based synthetic X-ray image generation

Abstract

Journal

Organization

Cited Papers

Citing Papers