课程摘要：Harmonic mappings are extensively used in geometry processing applications to produce visually appealing deformations. We establish the sufficient and necessary conditions for a harmonic planar mapping to have bounded distortion. Our key observation is that these conditions relate solely to the boundary behavior of the mapping. This leads to an efficient and accurate algorithm that supports handle-based interactive shape-and-image deformation and is demonstrated to outperform other state-of-the-art methods. The particular structure of harmonic mappings further allows efficient shape interpolation. Given the closed-form expressions for the interpolants, our interpolation algorithm runs embarrassingly in parallel and is orders of magnitude faster than state-of-the-art methods due to its simplicity, yet it produces mappings that are superior to those existing techniques due to guaranteed bounds on geometric distortions.
课程题目：Proximal Algorithms for Optimization in Computer Graphics
授课讲者：邓柏林，Cardiff University
课程摘要：In recent years, proximal algorithms have become a popular approaching for solving large-scale optimization problems. By introducing auxiliary variables and utilizing proximal operators, these algorithms allow parallel implementation with fast convergence to approximate solutions. In this talk, we will learn about applications of proximal algorithms in geometry processing and physical simulation. Examples include local-global solvers for unconstrained optimization, and alternating direction method of multipliers (ADMM) for constrained optimization. The talk will also present an acceleration technique that improves their convergence to high-accuracy solutions.
讲者简介：Bailin Deng is a lecturer at the School of Computer Science and Informatics of Cardiff University. His main research interest is in geometry processing techniques and computational design tools, especially in the context of freeform architectural design and digital fabrication. Previously he was a lecturer in the School of Engineering and Computer Science at University of Hull (2015-2017), and a postdoctoral researcher in the Computer Graphics and Geometry Laboratory at EPFL (2012-2015). He obtained his PhD degree in mathematics from Vienna University of Technology in 20111, and holds a master’s degree in computer science and a bachelor’s degree in computer software from Tsinghua University.
课程摘要：This course introduces the current state-of-the-art in image and video retargeting and describes important ideas and technologies that have influenced the recent work. Retargeting has retained in the front rank of most widely-used digital media processing techniques for a long time. It is the process of adapting an image or video from one screen resolution to another to fit different displays, for example, when watching a wide screen movie on a normal television screen or a mobile device. As there has been considerable work done in this field already, this course provides an overview of the techniques. It is meant to be a starting point for new research in the field. We include explanations of basic terms and operators, as well as the basic workflow of the different methods.
课程摘要：The light field (LF) image not only records the accumulated light intensity of the scene, but also implicitly encodes the three-dimensional geometry information of the scene. LF images facilitate/allow a wide range of interesting applications, e.g., image editing/post-capture refocus, 3-D reconstruction, etc. Moreover, the LF is considered as a promising paradigm for immersive 3-D telepresence, which can be widely applied to entertainment, telerehabilitation, business, etc., to significantly reform human work and life styles.
Recent advances in commercial hand-held LF cameras make the convenient acquisition of LF images in a single snapshot possible, which dramatically boosts LF-based research and applications. In this talk, I will introduce our recent progress on learning-based LF image processing, including LF image denoising/compression/depth estimation, and high-fidelity LF image reconstruction (spatial/angular super-resolution).
讲者简介：Junhui Hou received the B.Eng. degree in information engineering (Talented Students Program) from the South China University of Technology, Guangzhou, China, in 2009, the M.Eng. degree in signal and information processing from Northwestern Polytechnical University, Xian, China, in 2012, and the Ph.D. degree in electrical and electronic engineering from the School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, in 2016.
He has been an Assistant Professor with the Department of Computer Science, City University of Hong Kong, since 2017. His research interests fall into the general areas of visual signal processing, such as adaptive image/video representations and analysis (RGB/depth/light field/hyperspectral), static/dynamic 3-D geometry representations and processing (mesh/point cloud/MoCap), and discriminative modeling for lustering/classification.
Dr. Hou was the recipient of several prestigious awards, including the Chinese Government for Outstanding Self-Financed Students Abroad, China Scholarship Council in 2015), and the Early Career Award from the Hong Kong Research Grants Council in 2018. He serves/served as an Associate Editor for The Visual Computer, an Area Editor for Signal Processing: Image Communication, and the Guest Editor for the Journal of Visual Communication and Image Representation. He is an Area Chair of ACM International Conference on Multimedia (ACM MM) 2019. He was/is also involved in the organization of some international conferences, such as the local arrangement chair of the 26th Pacific Conference on Computer Graphics and Applications (PG), 2018 and the publication co-chairs of IEEE International Conference on Visual Communication and Image Processing (IEEE VCIP), 2020.
课程题目：Non-uniform Extraordinary Points: Geometry and Approximation
授课讲者：李新，中国科学技术大学
课程摘要：Extraordinary points (EPs) are necessary to construct arbitrary topological spline surfaces which play a central role in geometric modeling and isogeometric analysis. However, if the knot intervals are different for the edges connecting an extraordinary point, there are two main challenges for the non-uniform EPs. The first one is that the blending functions for extraordinary points of the existing methods can have two local maxima. And the second one is that the approximation order around the extraordinary points is not optimal. This talk will review the existing approaches for these problems and develops new refinement rules to handle the two problems, which can yield a G1 extraordinary point with better shape quality and optimal convergence rates.
课程题目：Towards Real-time Simulation of Deformable Objects
授课讲者：刘天添，微软亚洲研究院
课程摘要：We propose a new method for physics-based simulation supporting many different types of hyperelastic materials from mass-spring systems to three-dimensional finite element models, pushing the performance of the simulation towards real-time. Previous methods such as Position Based Dynamics are fast but support only limited selection of materials. Simulation of general types of materials currently relies on Newton's method, which is slow, even with only one iteration per timestep.
In this talk, we start from simple material models such as mass-spring systems or as-rigid-as-possible materials. We express the widely used implicit Euler time integration as an energy minimization problem and introduce auxiliary projection variables as extra unknowns. After our reformulation, the minimization problem becomes linear in the node positions, while all the non-linear terms are isolated in individual elements. We then show that our reformulation can be interpreted as a quasi-Newton method. This insight enables very efficient simulation of a large class of hyperelastic materials, including the Neo-Hookean, spline-based materials, and others. The quasi-Newton interpretation also allows us to leverage ideas from numerical optimization. In particular, we show that our solver can be further accelerated using L-BFGS updates (Limited-memory Broyden-Fletcher-Goldfarb-Shanno algorithm). Our final method of simulating deformable objects is typically more than 10 times faster than one iteration of Newton's method without compromising quality. In fact, our result is often more accurate than the result obtained with one iteration of Newton's method. Our method is also easier to implement, implying reduced software development costs.
讲者简介：Tiantian Liu is an associate researcher at the Internet Graphics group in Microsoft Research Asia. His research interests include physically-based simulation and fast geometry processing algorithms. Before joining MSRA, he obtained his Ph.D. degree in Computer and Information Science at the University of Pennsylvania with Prof. Ladislav Kavan. He also had a Master's degree in Computer Graphics and Game Technology at the University of Pennsylvania, and a Bachelor's degree in Computer Science and Technology at Zhejiang University.
课程摘要：Visual object tracking is challenging as target objects often undergo significant appearance changes. In this talk, I will present our work on how to best exploit deep regression networks to improve tracking accuracy and robustness. First, I will introduce our TPAMI 2019 (ICCV 2015) work, in which we adaptively learn correlation filters on hierarchical convolutional layers to precisely locate targets. Second, I will present our ECCV 2018 work, where we propose a novel shrinkage loss to train deep regression networks for visual tracking. Last, I will present two CVPR 2019 works, where we respectively learn a target-aware regression model and an unsupervised regression model for visual tracking. Extensive experimental results on largescale benchmark datasets show that the proposed algorithms perform favorably against state-of-the-art methods.
讲者简介：许威威，现任浙江大学CAD&C国家重点实验室百人计划研究员，曾任日本立命馆大学博士后，微软亚洲研究院网络图形组研究员, 杭州师范大学浙江省钱江学者特聘教授。主要研究兴趣为三维重建、物理仿真、3D打印和虚拟现实，发表高水平论文近70余篇，其中顶级期刊ACM Transactions on Graphics论文17篇，研究成果在三维扫描仪和人体三维重建等企业产品中得到应用。2014年获国家自然科学基金优秀青年基金资助。