TRACE develops, implements, and tests computer vision algorithms needed to build self-driving cars. As the safety of the public and of drivers and their passengers is at stake, the bar for the effectiveness and efficiency of these algorithms has to be put extraordinarily high. The functionalities under study include the extraction of stereo depth, the detection of other traffic agents, their tracking and re-identification, the recognition of lane markings and traffic signs, the detection of free space, the one-line application of structure-from-motion, and much more. The implementation of these functionalities involves optimized GPU processing. Candidate components are extensively tested on Toyota’s self-driving car prototypes.
Research topics
Videos
Research Papers
Dengxin Dai, Luc Van Gool
Dark Model Adaptation: Semantic Image Segmentation from Daytime to Nighttime
International Conference on Intelligent Transportation Systems (ITSC), November 4-7, 2018, Maui, Hawaii, USA
Simon Hecker, Dengxin Dai, Luc Van Gool
End-to-End Learning of Driving Models with Surround-View Cameras and Route Planners
European Conference on Computer Vision (ECCV), Munich, Germany, 2018
Christos Sakaridis, Dengxin Dai, Simon Hecker, Luc Van Gool
Model Adaptation with Synthetic and Real Data for Semantic Dense Foggy Scene Understanding
European Conference on Computer Vision (ECCV), Munich, Germany, 2018
Simon Hecker, Dengxin Dai, Luc Van Gool
Failure Prediction for Autonomous Driving
IEEE Intelligent Vehicles Symposium, Changshu, China, 2018
Yuhua Chen, Wen Li, Christos Sakaridis, Dengxin Dai, Luc Van Gool
Domain Adaptive Faster R-CNN for Object Detection in the Wild
Conference on Computer Vision and Pattern Recognition, Salt Lake City, USA, 2018
Arun Balajee Vasudevan, Dengxin Dai, Luc Van Gool
Object Referring in Videos with Language and Human Gaze
Conference on Computer Vision and Pattern Recognition, Salt Lake City, USA, 2018
Jonas Heylen, Seppe Iven, Bert De Brabandere, Jose Oramas, Luc Van Gool, Tinne Tuytelaars
From Pixels to Actions: Learning to Drive a Car with Deep Neural Networks
IEEE Winter Conf. on Applications of Computer Vision (WACV) 2018
Davy Neven, Bert De Brabandere, Stamatios Georgoulis, Marc Proesmans, Luc Van Gool
Towards End-to-End Lane Detection: an Instance Segmentation Approach
29th IEEE Intelligent Vehicles Symposium, Chang Shu, China, 2018
Arun Balajee Vasudevan, Dengxin Dai, Luc Van Gool
Object Referring in Visual Scene with Spoken Language
IEEE Winter Conf. on Applications of Computer Vision (WACV) 2018
Sakaridis C., Dai D., Van Gool L.
Semantic Foggy Scene Understanding with Synthetic Data
International Journal of Computer Vision, 2018
Manen S., Gygli M., Dai D., Van Gool L.
PathTrack: Fast Trajectory Annotation with Path Supervision
International Conference on Computer Vision, Venice, Italy, October 2017
De Brabandere B., Neven D., Van Gool L.
Semantic Instance Segmentation for Autonomous Driving
in Deep Learning for Robotic Vision Workshop (CVPR 2017), Hawaii, USA, July 2017 ( Best paper award )
Neven D., De Brabandere B., Georgoulis S., Proesmans M., Van Gool L.
Fast Scene Understanding for Autonomous Driving
in Deep Learning for Vehicle Perception Workshop (IV 2017), Redondo Beach, US, June 2017
Li W., Xu Z., Xu D., Dai D., Van Gool L.
Domain Generalization and Adaptation using Low Rank Exemplar SVMs
IEEE Transactions on Pattern Analysis and Machine Intelligence, May 2017
De Brabandere B., Jia X., Tuytelaars T., Van Gool L.
Dynamic filter networks
30th annual conference on neural information processing systems - NIPS 2016, 9 pp., December 4-9, 2016, Barcelona, Spain
Jia X., De Brabandere B., Tuytelaars T., Van Gool L.
Dynamic filter networks for predicting unobserved views
Proceedings ECCV 2016 workshops, 2 pp., October 8-16, 2016, Amsterdam, The Netherlands
Davy Neven won the third prize in the lane detection challenge at the Vision Based Autonomous Driving Challenge in Hawaii, July 2017.
Luc Van Gool, Distinguished Researcher Award at ICCV in Venice, Italy in October 2017