A Study on Computer Vision Techniques for Self-driving Cars

摘要

Self-driving cars have become inevitable to be present in a near future. A big number of large companies, startups and research groups have been working for years to fulfil the vision of an absolute unmanned transportation system. These systems have the capacity to model how our future societies and livelihood will be shaped. The Utopian dream may still be years away, but current researchers have been shaping up for this tomorrow little by little. From a rise and advances in the field of deep learning, there has been a major push received, especially in the field of computer vision and its possibly uncountable applications. Autonomous vehicles have grown a lot over the past decade from development of better computer vision algorithms for solving common driving tasks, to coming up with large datasets which support training of such systems. In this paper, we have aimed to give a brief introduction to research trends being followed in the overlapping areas of self-driving cars and computer vision. Some state of the art algorithms for solving some common problems which an autonomous system can be benefitted from, such as object detection and semantic scene segmentation, are also discussed.