Computer Graphics Laboratory

Course Topics

This seminar covers advanced topics in visual computing, including both seminal research papers as well as the latest research results. Topics include data-driven modeling and animation, image based rendering, real-time vision & graphics, physical simulation and generative models, visual perception, computational photography, video synthesis, and others.

Course Setup

Every participant has to present one of the papers in the list below. Additionally, you are required to read the paper that is presented in class beforehand and participate in a discussion during the seminar. An assistant will provide support when preparing the slides and in case technical questions arise.

The projector seems to have an aspect ratio of 16:10. Note, that it is not allowed to use the original slides from the author.

Learning Objectives

The goal is to get an in-depth understanding of actual problems and research topics in the field of visual computing as well as improve presentations and critical analysis skills.

Prerequisites

The "Visual Computing", "Introduction to Computer Graphics" and "Computer Vision I" courses are recommended, but not mandatory.

Administration

Schedule

Date	Name	Topic	Supervisor
Sept. 17	Lubor Ladicky	Introduction
Sept. 24	Lubor Ladicky	Example Talk, Common Potential Issues with Talks
Oct. 1	Himanshu Jain	Palette-Based Photo Recoloring	Yagiz Aksoy
Oct. 1	Hironori Yoshida	Microstructures to Control Elasticity in 3D Printing	Christian Schumacher
Oct. 8	Marcel Lüdi	Time-lapse Mining from Internet Photos	Changil Kim
Oct. 8	Simon Kallweit	Perceptually Based Downscaling of Images	Cengiz Öztireli
Oct. 15	Thomas Müller	Adaptive Rendering Based on Weighted Local Regression	Fabrice Rousselle
Oct. 15	Carlota Soler	Dynamic 3D Avatar Creation from Hand-held Video Input	Fabio Maninchedda
Oct. 22	Dominik Borer	High-Quality Capture of Eyes	Pascal Bérard
Oct. 22	Irene Baeza	A Computational Approach for Obstruction-Free Photography	Kaan Yücer
Oct. 29	Samuel Oberholzer	Data-Driven Color Manifolds	Marcel Lancelle
Oct. 29	Raphael Stadler	Rent3D: Floor-Plan Priors for Monocular Layout Estimation	Pablo Speciale
Nov. 5	Daniel Keyes	Real-Time Shading-Based Refinement for Consumer Depth Cameras	Chenglei Wu
Nov. 5	Delio Vicini	Sampling Based Scene-Space Video Processing	Jean-Charles Bazin
Nov. 12	Jonathan Forman	Light Field Reconstruction Using Sparsity in the Continuous Fourier Domain	Changil Kim
Nov. 12	Jan Wolf	Texture Mapping Real-World Objects with Hydrographics	Olga Diamanti
Nov. 19	Li-yi Liu	Fully Convolutional Networks for Semantic Segmentation	Lubor Ladicky
Nov. 19	Sandro Lombardi	3D Scanning Deformable Objects With a Single RGBD Sensor	Martin Oswald

Seminar Papers

• Sampling Based Scene-Space Video Processing
• Time-lapse Mining from Internet Photos
• Perceptually Based Downscaling of Images
• Dehazing using Color-Lines
• Data-Driven Color Manifolds
• An L1 Image Transform for Edge-Preserving Smoothing and Scene-Level Intrinsic Decomposition
• A Computational Approach for Obstruction-Free Photography
• Light Field Reconstruction Using Sparsity in the Continuous Fourier Domain
• Homogeneous Codes for Energy-Efficient Illumination and Imaging
• High-Quality Capture of Eyes
• Real-Time Shading-Based Refinement for Consumer Depth Cameras
• Real-Time High-Fidelity Facial Performance Capture
• Adaptive Rendering Based on Weighted Local Regression
• Image-Based Relighting Using Neural Networks
• Microstructures to Control Elasticity in 3D Printing
• Texture Mapping Real-World Objects with Hydrographics
• DynamicFusion: Reconstruction and Tracking of Non-rigid Scenes in Real-Time
• Palette-Based Photo Recoloring
• Category-Specific Object Reconstruction From a Single Image
• Robust Reconstruction of Indoor Scenes
• Fully Convolutional Networks for Semantic Segmentation
• Rent3D: Floor-Plan Priors for Monocular Layout Estimation
• Efficient Visual-Inertial Navigation using a Rolling-Shutter Camera with Inaccurate Timestamps
• Dynamic 3D Avatar Creation from Hand-held Video Input
• 3D Scanning Deformable Objects With a Single RGBD Sensor