This paper presents a confidence-based distributional framwork that can efficiently estimate the 6DOF poses of multiple
objects from a single RGB image in real time or accurately estimate objects in individual classes.
This paper proposes forming adaptable patches according to the content of the input image and collected photos. The numbers and shapes of these patches are dynamically adjusted by multilabel optimization.
In this paper, we propose a novel tangible interface to guide a user assembling the components in an intuitively way. And most of the users give positive responses to our prototype system in the user evaluation.
This paper presents a novel approach for extracting foreground objects from an image. This novel method integrates efficient image completion and graph labeling techniques into a powerful framework.
The goal of this paper is to efficiently represent motion capture data and retain few reconstruction errors as well. We propose a novel segmentation and indexing method, called repeated motion analysis (RMA).
Our goal is to synthesize realistic facial animation. To achieve the goal, we research on advanced face modeling, facial motion analysis/capture, model deformation and realistic rendering.
This research project is to tackle problems in character motion synthesis (from motion transition, motion blending, motion adjustment, to motion compression). We're also interested in motion effects on characters' appearance.