Jan Kautz

Latest

• GAvatar: Animatable 3D Gaussian Avatars with Implicit Mesh Learning
• 3D Reconstruction with Generalizable Neural Fields using Scene Priors
• Convolutional State Space Models for Long-Range Spatiotemporal Modeling
• Generalizable One-shot 3D Neural Head Avatar
• Zero-shot Pose Transfer for Unrigged Stylized 3D Characters
• CoordGAN: Self-Supervised Dense Correspondences Emerge from GANs
• FreeSOLO: Learning to Segment Objects without Annotations
• GroupViT: Semantic Segmentation Emerges from Text Supervision
• Learning Contrastive Representation for Semantic Correspondence
• Self-Supervised Object Detection via Generative Image Synthesis
• Learning to Track Instances without Video Annotations
• Weakly-Supervised Physically Unconstrained Gaze Estimation
• Learning Continuous Environment Fields via Implicit Functions
• Online Adaptation for Consistent Mesh Reconstruction in the Wild
• Optical gaze tracking with spatially-sparse single-pixel detectors
• Self-supervised Single-view 3D Reconstruction via Semantic Consistency
• Self-Supervised Viewpoint Learning From Image Collections
• Joint-task Self-supervised Learning for Temporal Correspondence
• Few-Shot Adaptive Gaze Estimation
• Few-Shot Viewpoint Estimation
• Learning Propagation for Arbitrarily-structured Data
• Switchable Temporal Propagation Network
• Light-weight Head Pose Invariant Gaze Tracking
• Learning Affinity via Spatial Propagation Networks
• Dynamic Facial Analysis: From Bayesian Filtering to Recurrent Neural Network
• Mixed-primary factorization for dual-frame computational displays.