Millimeter-Wave Image Deblurring via Cycle-Consistent Adversarial Network

Millimeter-wave (MMW) imaging has a tangible prospect in concealed weapon detection for security checks. Typically, one-dimensional (1D) linear antenna array with mechanical scanning along perpendicular direction is employed MMW imaging. To achieve high-resolution imaging, the target under test needs to keep steady enough during process since slight movement can induce large phase variation systems, which will result blurred image. However, scenario of human body, sometimes it difficult meet this requirement, especially elderly. Such images would reduce accuracy weapons. In paper, we propose deblurring method based on cycle-consistent adversarial network (Cycle GAN). Specifically, Cycle GAN learn mapping between and focused ones. minimize effect shaking blur, introduce an identity loss. Moreover, mean squared error loss (MSE loss) utilized stabilize training, so as obtain more refined deblurred results. The experimental results demonstrate that proposed efficiently suppress blurring