CVPR Daily - Tuesday

Shuangjun’s picks of the day: Shuangjun Liu is a 3rd year PhD student of the Augmented Cognition Lab (ACLab) at the ECE Department of Northeastern University in Boston, USA. He enjoys traveling and photography, as well as figuring out things and exploring new territories. For these reasons, research work appeals to him a lot: “ I especially like to solve real-world problems. With rapid emerging works towards various real-world applications, computer vision and machine learning no doubt have become one of my favorite research directions. One of the intriguing topics for me is how to make a machine understand an image like a human .” “Currently, my research focus is centered around solving visual perception problems including human pose estimation under specific contexts with data limitation, such as medical care. In particular, I am interested in bringing domain knowledge into the deep learning structures through the use of physics-guided simulation as a means for data augmentation. I am glad to be at CVPR19 to learn great works and talk to the community. I am presenting two posters on behalf of my lab members at ACLab: " Infant Contact-less Non-Nutritive Sucking Pattern Quantification via Facial Gesture Analysis " (at workshop on Augmented Human) and " Introduction to Indoor GeoNet " (at workshop on 3D Scene Understanding for Vision, Graphics, and Robotics)” Orals: 1.1A-4 Kervolutional Neural Networks 1.1B-76 DeepSDF Learning Continuous Signed Distance Functions for Shape Repr. 1.2B-121 Progressive Pose Attention Transfer for Person Image Generation Posters: 1.2-6 Universal domain adaptation 1.2-92 Exploiting Temporal Context for 3D Human Pose Estimation in the Wild 1.2-140 Dense Intrinsic Appearance Flow for Human Pose Transfer For today, Tuesday 18 2 Shuangjun’s Picks DAILY CVPR Tuesday