Abstract In image based rendering methods, viewing geometry is used to ensure a physically valid image transforms. We compare and contrast methods based on affine and pro-jective viewing geometry and with and without an explicit 3D object... more
Abstract In image based rendering methods, viewing geometry is used to ensure a physically valid image transforms. We compare and contrast methods based on affine and pro-jective viewing geometry and with and without an explicit 3D object structure representation. 1 ...
Research Interests:
... presented a new type of navigation map that combines appearance information with sparse geometric information about the ... In the future we want to improve the localization algorithm byembedding a statistical model of the ... 3D... more
... presented a new type of navigation map that combines appearance information with sparse geometric information about the ... In the future we want to improve the localization algorithm byembedding a statistical model of the ... 3D model construction using range and image data. ...
Page 1. 1 A Survey and Tutorial on Shape and Reflectance Estimation Techniques Dana Cobzas1, Neil Birkbeck1, Martin Jagersand1 and Peter Sturm2 1Computer Science, University of Alberta, Canada 2INRIA Rhone-Alpes, France April 26, 2006... more
Page 1. 1 A Survey and Tutorial on Shape and Reflectance Estimation Techniques Dana Cobzas1, Neil Birkbeck1, Martin Jagersand1 and Peter Sturm2 1Computer Science, University of Alberta, Canada 2INRIA Rhone-Alpes, France April 26, 2006 DRAFT Page 2. 2 Abstract ...
We present a system designed for capturing 3D shape and appearance of objects of different sizes from 2D images. The capture system requires only a consumer camera, a rotating table and a calibration pattern. The whole modeling process... more
We present a system designed for capturing 3D shape and appearance of objects of different sizes from 2D images. The capture system requires only a consumer camera, a rotating table and a calibration pattern. The whole modeling process takes about 5-10 ...
Research Interests:
ABSTRACT The estimation of body composition (i.e., proportions of muscle and fat tissues) in cancer patients has important clinical and research applications. In particular, chemotherapy drug dosage is determined after taking into account... more
ABSTRACT The estimation of body composition (i.e., proportions of muscle and fat tissues) in cancer patients has important clinical and research applications. In particular, chemotherapy drug dosage is determined after taking into account the muscle and fat proportions in the patient's body. Recently, there has been considerable interest in studying the correlation between survival and body composition in cancer patients. We propose a fully automated framework for segmentation and quantification of muscle and fat tissues in thoracic CT images. A novel approach based on statistical deformation model (SDM) constrained deformable registration using the finite element method (FEM) is proposed. We obtained very good segmentation results with Jaccard scores of 94.95% for muscle and 94.82% for fat tissues respectively on a large data set of 116 thoracic CT images.
We present a novel variational formulation of discrete deformable registration as the minimization of a convex energy functional that involves diffusion regularization. We show that a finite difference solution (FD) of the variational... more
We present a novel variational formulation of discrete deformable registration as the minimization of a convex energy functional that involves diffusion regularization. We show that a finite difference solution (FD) of the variational formulation is equivalent to a continuous-valued Gaussian Markov random field (MRF) energy minimization formulation previously proposed as the random walker deformable registration method. A computationally efficient solution using the finite element method (FEM) method has been proposed to solve the variational minimization problem. Our proposed method obtained competitive results when compared with 14 other deformable registration methods on the CUMC12 MRI dataset.
Research Interests:
Abstract Having an accurate model of the lights in a scene is important for many applications. Augmented reality requires light information for seamlessly com-posing computer generated objects into a real scene. Similarly, computer vision... more
Abstract Having an accurate model of the lights in a scene is important for many applications. Augmented reality requires light information for seamlessly com-posing computer generated objects into a real scene. Similarly, computer vision algorithms such as ...
Research Interests: Computer Vision, Image Processing, Numerical Method, Brain Tumor, Computer Vision and Pattern Recognition, and 9 moreDecode-And-Forward, Theory and Practice, Anisotropic Diffusion, Parabolic Wave Equation, Second Order, Acoustic Diffusion Equation Model, Finite Difference Method, Differential equation, and Tumor Growth
Proceedings of the 1003 IEEE International Conferrocr on Robotics & AUfOmdoO Taipei, Taiwan, September 14-19, 1003 ... Predictive Display Models for Tele-Manipulation from Uncalibrated Camera-capture of Scene Geometry... more
Proceedings of the 1003 IEEE International Conferrocr on Robotics & AUfOmdoO Taipei, Taiwan, September 14-19, 1003 ... Predictive Display Models for Tele-Manipulation from Uncalibrated Camera-capture of Scene Geometry and Appearance ... Keith Yerex, Dana Cobzas and ...
Research Interests:
... We propose a supervised variational method that incorporates additional appearance priors to better disambiguate the tumor from the surrounding deformed brain tissue. The formulation extends the Chan-Vese region-based segmentation... more
... We propose a supervised variational method that incorporates additional appearance priors to better disambiguate the tumor from the surrounding deformed brain tissue. The formulation extends the Chan-Vese region-based segmentation model14 in a similar way to texture ...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Robot tele-operation is significantly degraded by delays in the operator visual feedback. We present a cylindrical image-depth model that can be easily acquired at the remote work site and then used to support both robot localization and... more
Robot tele-operation is significantly degraded by delays in the operator visual feedback. We present a cylindrical image-depth model that can be easily acquired at the remote work site and then used to support both robot localization and predictive display. We present an ...
Research Interests:
ABSTRACT We propose a novel template-based multi-region segmentation method using a finite element method (FEM) deformation model with diffusion-based regularization. Our proposed method is computationally more efficient than the... more
ABSTRACT We propose a novel template-based multi-region segmentation method using a finite element method (FEM) deformation model with diffusion-based regularization. Our proposed method is computationally more efficient than the traditional template-based segmentation methods that use non-parametric or B-spline based deformation models, as it significantly reduces the number of degrees of freedom (DOF) associated with the energy minimization that arises in template-based segmentation. Further, like all template-based segmentation approaches our method is able to preserve topology of the initial regions of interest (ROIs) defined in the template, which is very useful for segmentation of anatomical structures. Segmentation results on medical images with various anatomical structures show that the proposed method improves computational efficiency without compromising segmentation accuracy.
Research Interests:
Abstract We present a texturing approach for image-based modeling and render-ing, where instead of texturing from one (or a blend of a few) sample images, new view-dependent textures are synthesized by modulating a differential texture... more
Abstract We present a texturing approach for image-based modeling and render-ing, where instead of texturing from one (or a blend of a few) sample images, new view-dependent textures are synthesized by modulating a differential texture basis. The texture basis contains ...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
ABSTRACT Existing scene flow approaches mainly focus on two-frame stereo-pair configurations and reconstruct an image-based representation of scene flow. Instead, we propose a variational formulation of scene flow relative to a coarse... more
ABSTRACT Existing scene flow approaches mainly focus on two-frame stereo-pair configurations and reconstruct an image-based representation of scene flow. Instead, we propose a variational formulation of scene flow relative to a coarse proxy geometry, which is better suited for many views. Furthermore, a linear basis is used to represent temporal surface flow, allowing for longer-range temporal correspondence with fewer variables. Our formulation takes known proxy motion into account (e.g, if the proxy is a tracked human subject), which enables 3D trajectory reconstruction when only a single view is available. Additionally, through the appropriate proxy and basis, our framework generalizes existing approaches for scene flow, optic-flow, and two-frame stereo. We illustrate results on real-data for both static and moving proxy surfaces over several frames.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests: Engineering, Algorithms, Diffusion Tensor Imaging, Medical Image Analysis, Clinical Practice, and 10 moreHumans, Glioma, Neoplasm Invasiveness, Image Enhancement, Reproducibility of Results, Sensitivity and Specificity, Brain Neoplasms, Euclidean Distance, Medical Image Computing, and Geodesic Distance
Research Interests:
Research Interests:
The problem of capturing real world scenes from images and then rendering them is of significant interest in both computer vision and graphics research. It has important applications in architectural reconstruction and visualization,... more
The problem of capturing real world scenes from images and then rendering them is of significant interest in both computer vision and graphics research. It has important applications in architectural reconstruction and visualization, virtual museums, landscape scene capturing for robot ...
While many automatic medical image segmentation al-gorithms have been proposed in the literature, they have not made it into clinical use. Two reasons are that algorithms still fail on challenging data, and that doctors wish to re-tain... more
While many automatic medical image segmentation al-gorithms have been proposed in the literature, they have not made it into clinical use. Two reasons are that algorithms still fail on challenging data, and that doctors wish to re-tain direct control over segmentation decisions. Thus in ...
1 Introduction Virtual Heritage (VH) is concerned with the digital repre-sentation and presentation of cultural artifacts along with the stories and facts that accompany them. In this work, we present an inexpensive, easy to use,... more
1 Introduction Virtual Heritage (VH) is concerned with the digital repre-sentation and presentation of cultural artifacts along with the stories and facts that accompany them. In this work, we present an inexpensive, easy to use, Image-Based Modeling and Rendering (IBMR) system for ...