Towards explainable semantic segmentation for autonomous driving systems by multi-scale variational attention

Explainable autonomous driving systems (EADS) are emerging recently as a combinatory field of explainable artificial intelligence (XAI) and vehicular automation (VA). EADS explains events, ambient environments, and engine operations of an autonomous driving vehicular, and it also delivers explainable results in an orderly manner. Explainable semantic segmentation (ESS) plays an essential role in building EADS, where it offers visual attention that helps the drivers to be aware of the ambient objects irrespective if they are roads, pedestrians, animals, or other objects. In this paper, we propose the first ESS model for EADS based on the variation autoencoder (VAE), and it uses the multiscale first-order derivatives between the latent space and the encoder layers to capture the curvatures of the neurons’ responses. Our model is termed as Mgrad2VAE and is bench-marked on the SYNTHIA dataset, where it outperforms the recent deep models in terms of image segmentation metrics.