Bin Wang is currently a full-time researcher of Advanced Innovation Center for Future Visual Entertainment in Beijing Film Academy. She earned her Ph.D in November 2012 from Beihang University, advised by Prof. Shuling Dai. From Aug. 2010 to Aug. 2012, she experienced two-year’s wonderful life in the Sensorimotor System Lab at University of British Columbia, Canada, working with Prof. Dinesh K. Pai and Prof. François Faure, from University Joseph Fourier, Grenoble, France. She has been a Research Fellow from April 2013 to April 2014 in National University of Singapore, supervised by Prof. Kangkang Yin. From 2014 to 2016, she did her postdoctoral researcher at Visual Computing Research Center, leading by Prof. Hui Huang.
Quickly discover relevant content by filtering publications.
This paper proposes a novel three-way coupling framework to simulate the surface-tension-dominant contact between rigid and fluid, which using a Lagrangian surface membrane to handle the interactions between bodies and fluid.
This paper proposes an unsupervised Co-part segmentation approach, which leverages shape correlation information between different frames in the video to achieve semantic part segmentation. We have designed a novel network structure which achieves self-supervision through a dual procedure of part-assembly to form a closed loop with part-segmentation. Additionally, we have developed several new loss functions that ensure consistent, compact and meaningful part segmentation and the intermediate transformations with clear explainable physical meaning.
This paper presents a new method for estimating nonlinear constitutive models from trajectories of surface data. The key insight is to have a parametric material correction model learn the error of the elastic and damping properties of a nominal material.
This paper presents a versatile, numerical approach to simulating various magnetic phenomena using a level-set method, which contains a novel two-way coupling mechanism between a magnetic field and magnetizable objects.
In this paper, we propose an example-driven camera controller which can extract camera behaviors from an example film clip and re-apply the extracted behaviors to a 3D animation, through learning from a collection of camera motions.
We present a data-driven method that can capture deformation of generic soft objects in high fidelity with low-cost depth sensors; and estimate plausible deformation parameters from these pure kinematic motion trajectories, without requiring any force-displacement measurements as is common in traditional methods. Using the learned deformation models, new motion and deformation can be synthesized at interactive rates to respond to dynamic perturbations or satisfy user-specified constraints.
We have presented a two-way simulation and control framework for soft characters with inherent skeletons. We propose a novel pose-based plasticity model that extends the corotated linear elasticity model to achieve large skin deformation around joints. We further reconstruct controls from reference trajectories captured from human subjects by augmenting a sampling-based algorithm.
The accuracy of intersection volume is important for plausible collision response. In this paper we have presented the first imagebased collision detection method that provides the controllability of intersection volume without explicit geometrical computation, and demonstrated its relevance for precise contact modeling. Its computation combines rasterization at moderate resolution with adaptive ray casting, which allows more precise contact modeling where needed and a reduced memory footprint