Visage Technologies
Face Tracking & Analysis


ECCV 2018: European Conference on Computer Vision

Sep 20, 2018

The ECCV conference is one of the world’s top three academic gatherings in the field of computer vision (along with ICCV and CVPR). The 15th European Conference on Computer Vision (ECCV) was held in Munich, Germany. ECCV 2018 runs to September 14, with various tutorials and workshops focused on specific topics. The main technical program of the conference runs September 10–12.

This year, over 3,200 researchers attended, doubling the number from two years ago to make this the largest ECCV ever. The conference also received a record-high 2439 main conference submissions and accepted 776 (59 orals, 717 posters, 31.8% acceptance rate). The ECCV organizing committee expanded their reviewer team to deal with the increasing paper volume. Four program co-chairs, 126 area chairs, and 1199 reviewers combined to deliver over 7,300 paper reviews. Our colleagues attended the conference to gain insight to latest trends and knowledge.

Best Paper category is Implicit 3D Orientation Learning for 6D Object Detection from RGB Images from researchers at the German Aerospace Center and Technical University of Munich. The authors propose a real-time RGB-based pipeline for object detection and 6D pose estimation. The novel 3D orientation estimation is based on a variant of the Denoising Autoencoder that is trained on simulated views of a 3D model using Domain Randomization.

Honorable Mentions were awarded to the paper Group Normalization from Facebook AI Research, which proposes a simple but effective alternative to batch normalization that can accelerate deep neural network training with small batch sizes; and GANimation: Anatomically-aware Facial Animation from a Single Image from Spain’s Institute of Robotics and Industrial Informatics and the Ohio State University, which introduces a GAN conditioning scheme based on Action Units (AU) annotations, which describe in a continuous manifold the anatomical facial movements defining a human expression.

The papers can be read here:



← back  
Go to top