The papers cover a vast array of topics including an idea on how to improve vision and language navigation and a new model that is learning to generate grounded visual captions without localized supervision.
“ECCV is one of the top computer vision conferences, and Georgia Tech’s strong showing reflects the large amount of amazing faculty and students working in this important area. One thing that stands out is not only the amount of papers accepted but the diversity of the topics as well,” said Zsolt Kira, ML@GT associate director, an assistant professor in the School of Interactive Computing (IC) and co-author of two accepted papers.
The conference also features 45 workshops, 16 tutorials, and networking opportunities with companies like Sony, Google, and Autodesk.
Dhruv Batra, an associate professor in IC and ML@GT will give a talk at the Commands for Autonomous Vehicles (C4AV) workshop. Judy Hoffman, an assistant professor in IC and ML@GT will give two invited talks – one at the 2020 ChaLearn Looking at People workshop: Fair Face Recognition and Analysis and the Adversarial Robustness in the Real World workshop. ML@GT executive director Irfan Essa served as a 2020 area chair.
The conference, which meets every two years, is pivoting to a virtual event set for August 23-28. For more information on Georgia Tech’s accepted work, visit https://b.gatech.edu/3gOYb5a.