We investigate the viability of a variational U-Net architecture for denoising of single-channel audio data. Deep network speech enhancement systems commonly aim to estimate filter masks, or opt to ...

In this paper, we propose a Geometric Model-Free Approach with a Particle Filter (GMFA-PF) through the use of automotive LiDAR for real-time tracking of moving objects within an urban driving ...