Semantic image segmentation for autonomous driving is a challenging task due to its requirement for both effectiveness and efficiency. Recent developments in deep learning have demonstrated important performance boosting in terms of accuracy. In this paper, we present a comprehensive overview of the state-of-the-art semantic image segmentation methods using deep-learning techniques aiming to operate in real time so that can efficiently support an autonomous driving scenario. To this end, the presented overview puts a particular emphasis on the presentation of all those approaches which permit inference time reduction, while an analysis of the existing methods is addressed by taking into account their end-to-end functionality, as well as a comparative study that relies upon a consistent evaluation framework. Finally, a fruitful discussion is presented that provides key insights for the current trend and future research directions in real-time semantic image segmentation with deep learning for autonomous driving.

Drivable Area Segmentation and Lane Detection constitute crucial tasks for the Visual Perception system of an Autonomous Vehicle. The majority of the approaches dealing with these tasks are addressed as Semantic Segmentation problems using heavy deep learning models that become computationally expensive. In this paper, a dual-task lightweight model is proposed, which comprises a novel dual-task feature fusion mechanism allowing it to exploit global, high-level information while retaining useful low-level details for each task. This model excels not only in terms of accuracy but also achieves real-time performance by solving these two tasks in a multi-task fashion. Our comparative study which was conducted on the standard BDD100 K dataset shows that our proposed method compares favorably with the state-of-the-art offering an optimal trade-off between accuracy and efficiency.

Test-time adaptation enables a pre-trained model to update its weights during inference, in order to adapt to a target domain that has a different distribution from the source domain. This adaptation occurs without any supervision and often in a more challenging source-free setting where no data from the source domain are used. While test-time adaptation has received considerable attention for classification tasks, domain adaptation is equally important for other computer vision tasks, such as object detection. Many approaches consider a static target domain, which fails to simulate real-world conditions, where the target domain is non-stationary and the target distribution can gradually change over time. In this work, we focus on the continual test-time adaptation scenario, in which the target domain is continually changing over time. Leveraging the mean teacher framework for object detection, we stochastically restore a small part of the student’s weights to the source pre-trained model weights during adaptation. Additionally, we aim to enhance performance by using contrastive learning. After a consistent experimental work, it is shown that our proposed method compares favorably with the standard mean teacher approach.