Computer Vision News - May 2019

General Motors Research & Development. They also gained support from KPIT. The Computer Vision Center (CVC) in Barcelona and the Open Source Vision Foundation also host CARLA. Since CARLA can be run in the cloud, the challenge now has support from Amazon Web Services, in addition to Waymo, Uber, Audi Electronics Venture, Alpha Drive and EvalAI, who are also interested in the results of the challenge. All members of the consortium have an interest in autonomous driving. CARLA has evolved to include new functionalities based on feedback from participants coming from academia as well as from industry. Involving both sides of the community requires flexibility. They thought the best way of doing this was to propose a challenge. As people present different ideas, they can decide which ideas look more promising. Later, they can look at the results and see how the community responds. Behind the CARLA Challenge “The state-of-the-art of computer vision has been traditionally driven by challenges such as the PASCAL Visual Object Classes challenge, and the ImageNet Large Scale Visual Recognition Challenge. We work to make CARLA to play the same role for autonomous driving, at least when it comes to AI” explains Antonio. The autonomous driving community has several ways of approaching the problem of self-driving cars. Some take a more traditional approach based on pipelines of different modules, for example, doing perception based on computer vision. This involves object detection and semantic segmentation which rely on neural networks. The results are given to a control module to maneuver the car locally so that globally it can move from a starting point to a destination. However, this method requires annotating a lot of data to train the 34 Computer Vision News Challenge and Workshop - CVPR 2019 Challenge