interactive shape modeling and animation
The creation of digital geometric content is one of the fundamental tasks in computer graphics and is indispensible in many areas of our life. Most man-made or manufactured objects around us, as well as man-made landscapes, are modeled on a computer first. Geometric modeling is essential in engineering design and simulation, product prototyping and manufacturing, surgical planning, prosthesis design, architecture, geology, art, films and video games, and more. With the advance of modern fabrication technology such as 3D printing, digitally modeled shapes are increasingly easy to realize as physical objects.
The goal of our research is to lay theoretical foundations for shape modeling and develop practical algorithms and systems for digital shape creation, editing and animation. We seak to make 3D modeling accessible and effective for a wide range of users. On the algorithmic side, we are interested in fast yet versatile and powerful modeling and editing approaches that should help artists and designers be more productive and creative. We also look into intuitive interfaces for modeling and animation that can be used by non-experts. Our objective is to make 3D geometry content as ubiquitous and widely accessible as digital images and videos are today.
Our work touches on various aspects of shape modeling and animation, both in research and education. Representative topics: physically-plausible shape editing, variational sketch-based modeling, skinning deformation and animation, geometry cloning, efficient and robust numerical methods for shape modeling, volumetric modeling, expressive nonrealistic modeling, fabrication-aware geometric modeling. More info...
digital geometry processing
Digital 3D models, especially those acquired from real-world objects by scanning technologies, go through a digital geometry processing pipeline to prepare them for end applications. Processing operations include denoising and smoothing, resampling and mesh optimization, registration, parameterization, compression and more. We employ principles from discrete differential geometry, variational optimization and differential shape representations for a wide variety of geometry processing tasks. Our goal is to develop theoretical shape understanding and efficient and robust practical algorithms for delivering high-quality geometric data. Example topics include: 3D reconstruction, 3D scan analysis and segmentation, quad remeshing, retopology, global and local parameterization, inter-surface mapping. More info...
image and video processing
Image and video processing and manipulation can benefit from geometric approaches, whether applied to the spatial, temporal or color domain. We use differential representations and variational modeling for various applications in image and video editing, visual media retargeting, image-based rendering and more. For example, we design special warping functions for images and video that are optimized to preserve specific features, such as user-defined areas, salient content, faces, camera and object motion and temporal coherence. Our recent areas of research in this domain include reuse of visual data for creative design via guided texture synthesis or sub-pixel accurate transfer mapping, video-based 3D reconstruction, image and video retargeting for display devices of varying aspect ratios and color spaces, image-based scene navigation. More info...
The research at the Interactive Geometry Lab has been funded in part by the following sources: