Search For:

Displaying 1-28 out of 28 total
3D Scene Understanding by Voxel-CRF
Found in: 2013 IEEE International Conference on Computer Vision (ICCV)
By Byung-Soo Kim,Pushmeet Kohli,Silvio Savarese
Issue Date:December 2013
pp. 1425-1432
Scene understanding is an important yet very challenging problem in computer vision. In the past few years, researchers have taken advantage of the recent diffusion of depth-RGB (RGB-D) cameras to help simplify the problem of inferring scene semantics. How...
 
Efficient Human Pose Estimation from Single Depth Images
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Jamie Shotton,Ross Girshick,Andrew Fitzgibbon,Toby Sharp,Mat Cook,Mark Finocchio,Richard Moore,Pushmeet Kohli,Antonio Criminisi,Alex Kipman,Andrew Blake
Issue Date:December 2013
pp. 2821-2840
We describe two new approaches to human pose estimation. Both can quickly and accurately predict the 3D positions of body joints from a single depth image without using any temporal information. The key to both approaches is the use of a large, realistic, ...
 
Image Binarization for End-to-End Text Understanding in Natural Images
Found in: 2013 12th International Conference on Document Analysis and Recognition (ICDAR)
By Sergey Milyaev,Olga Barinova,Tatiana Novikova,Pushmeet Kohli,Victor Lempitsky
Issue Date:August 2013
pp. 128-132
While modern off-the-shelf OCR engines show particularly high accuracy on scanned text, text detection and recognition in natural images still remains a challenging problem. Here, we demonstrate that OCR engines can still perform well on this harder task a...
 
Decision tree fields
Found in: Computer Vision, IEEE International Conference on
By Sebastian Nowozin,Carsten Rother,Shai Bagon,Toby Sharp, Bangpeng Yao,Pushmeet Kohli
Issue Date:November 2011
pp. 1668-1675
This paper introduces a new formulation for discrete image labeling tasks, the Decision Tree Field (DTF), that combines and generalizes random forests and conditional random fields (CRF) which have been widely used in computer vision. In a typical CRF mode...
 
Efficient regression of general-activity human poses from depth images
Found in: Computer Vision, IEEE International Conference on
By Ross Girshick,Jamie Shotton,Pushmeet Kohli,Antonio Criminisi,Andrew Fitzgibbon
Issue Date:November 2011
pp. 415-422
We present a new approach to general-activity human pose estimation from depth images, building on Hough forests. We extend existing techniques in several ways: real time prediction of multiple 3D joints, explicit learning of voting weights, vote compressi...
 
A spatially varying PSF-based prior for alpha matting
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Christoph Rhemann, Carsten Rother, Pushmeet Kohli, Margrit Gelautz
Issue Date:June 2010
pp. 2149-2156
In this paper we considerably improve on a state-of-the-art alpha matting approach by incorporating a new prior which is based on the image formation process. In particular, we model the prior probability of an alpha matte as the convolution of a high-reso...
 
On detection of multiple object instances using hough transforms
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Olga Barinova, Victor Lempitsky, Pushmeet Kohli
Issue Date:June 2010
pp. 2233-2240
To detect multiple objects of interest, the methods based on Hough transform use non-maxima supression or mode seeking in order to locate and to distinguish peaks in Hough images. Such postprocessing requires tuning of extra parameters and is often fragile...
 
Surface stereo with soft segmentation
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Michael Bleyer, Carsten Rother, Pushmeet Kohli
Issue Date:June 2010
pp. 1570-1577
This paper proposes a new stereo model which encodes the simple assumption that the scene is composed of a few, smooth surfaces. A key feature of our model is the surface-based representation, where each pixel is assigned to a 3D surface (planes or B-splin...
 
Energy minimization for linear envelope MRFs
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Pushmeet Kohli, M. Pawan Kumar
Issue Date:June 2010
pp. 1863-1870
Markov random fields with higher order potentials have emerged as a powerful model for several problems in computer vision. In order to facilitate their use, we propose a new representation for higher order potentials as upper and lower envelopes of linear...
 
Dynamic Hybrid Algorithms for MAP Inference in Discrete MRFs
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Karteek Alahari, Pushmeet Kohli, Philip H.S. Torr
Issue Date:October 2010
pp. 1846-1857
In this paper, we present novel techniques that improve the computational and memory efficiency of algorithms for solving multilabel energy functions arising from discrete mrfs or crfs. These methods are motivated by the observations that the performance o...
 
Robust higher order potentials for enforcing label consistency
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Pushmeet Kohli, L'ubor Ladicky, Philip H. S. Torr
Issue Date:June 2008
pp. 1-8
This paper proposes a novel framework for labelling problems which is able to combine multiple segmentations in a principled manner. Our method is based on higher order conditional random fields and uses potentials defined on sets of pixels (image segments...
 
Reduce, reuse & recycle: Efficiently solving multi-label MRFs
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Karteek Alahari, Pushmeet Kohli, Philip H. S. Torr
Issue Date:June 2008
pp. 1-8
In this paper, we present novel techniques that improve the computational and memory efficiency of algorithms for solving multi-label energy functions arising from discrete MRFs or CRFs. These methods are motivated by the observations that the performance ...
 
Exact inference in multi-label CRFs with higher order cliques
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Srikumar Ramalingam, Pushmeet Kohli, Karteek Alahari, Philip H. S. Torr
Issue Date:June 2008
pp. 1-8
This paper addresses the problem of exactly inferring the maximum a posteriori solutions of discrete multi-label MRFs or CRFs with higher order cliques. We present a framework to transform special classes of multi-label higher order functions to submodular...
 
Dynamic Graph Cuts for Efficient Inference in Markov Random Fields
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Pushmeet Kohli, Philip H. S. Torr
Issue Date:December 2007
pp. 2079-2088
Abstract—In this paper we present a fast new fully dynamic algorithm for the st-mincut/max-flow problem. We show how this algorithm can be used to efficiently compute MAP solutions for certain dynamically changing MRF models in computer vision su...
 
P3&Beyond: Solving Energies with Higher Order Cliques
Found in: Computer Vision and Pattern Recognition, IEEE Computer Society Conference on
By Pushmeet Kohli, M. Pawan Kumar, Philip H. S. Torr
Issue Date:June 2007
pp. 1-8
In this paper we extend the class of energy functions for which the optimal ?-expansion and ??-swap moves can be computed in polynomial time. Specifically, we introduce a class of higher order clique potentials and show that the expansion and swap moves fo...
 
Effciently Solving Dynamic Markov Random Fields Using Graph Cuts
Found in: Computer Vision, IEEE International Conference on
By Pushmeet Kohli, Philip H. S. Torr
Issue Date:October 2005
pp. 922-929
In this paper we present a fast new fully dynamic algorithm for the st-mincut/max-flow problem. We show how this algorithm can be used to efficiently compute MAP estimates for dynamically changing MRF models of labelling problems in computer vision, such a...
 
Efficient Energy Minimization for Enforcing Label Statistics
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Yongsub Lim,Kyomin Jung,Pushmeet Kohli
Issue Date:September 2014
pp. 1893-1899
Energy minimization algorithms, such as graph cuts, enable the computation of the MAP solution under certain probabilistic models such as Markov random fields. However, for many computer vision problems, the MAP solution under the model is not the ground t...
 
Image Segmentation UsingHigher-Order Correlation Clustering
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Sungwoong Kim,Chang D. Yoo,Sebastian Nowozin,Pushmeet Kohli
Issue Date:September 2014
pp. 1761-1774
In this paper, a hypergraph-based image segmentation framework is formulated in a supervised manner for many high-level computer vision tasks. To consider short- and long-range dependency among various regions of an image and also to incorporate wider sele...
 
Relating Things and Stuff via ObjectProperty Interactions
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Min Sun,Byung-soo Kim,Pushmeet Kohli,Silvio Savarese
Issue Date:July 2014
pp. 1370-1383
In the last few years, substantially different approaches have been adopted for segmenting and detecting “things” (object categories that have a well defined shape such as people and cars) and “stuff” (object categories which have an amorphous spatial exte...
 
Relating Things and Stuff via Object Property Interactions
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By Min Sun,Byung-soo Kim,Pushmeet Kohli,Silvio Savarese
Issue Date:October 2013
pp. 1
In the last few years, substantially different approaches have been adopted for segmenting and detecting "things" (object categories that have a well defined shape such as people and cars) and "stuff" (object categories which have an am...
 
Associative Hierarchical Random Fields
Found in: IEEE Transactions on Pattern Analysis and Machine Intelligence
By L'ubor Ladicky,Chris Russell,Pushmeet Kohli,Philip H. S. Torr
Issue Date:August 2013
pp. 1
This paper makes two contributions: the first is the proposal of a new model - the associative hierarchical random field (AHRF), and a novel algorithm for its optimisation; the second is the application of this model to the problem of semantic segmentation...
 
Compressible Motion Fields
Found in: 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
By Giuseppe Ottaviano,Pushmeet Kohli
Issue Date:June 2013
pp. 2251-2258
Traditional video compression methods obtain a compact representation for image frames by computing coarse motion fields defined on patches of pixels called blocks, in order to compensate for the motion in the scene across frames. This piecewise constant a...
 
GeoF: Geodesic Forests for Learning Coupled Predictors
Found in: 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
By Peter Kontschieder,Pushmeet Kohli,Jamie Shotton,Antonio Criminisi
Issue Date:June 2013
pp. 65-72
Conventional decision forest based methods for image labelling tasks like object segmentation make predictions for each variable (pixel) independently [3, 5, 8]. This prevents them from enforcing dependencies between variables and translates into locally i...
 
A Principled Deep Random Field Model for Image Segmentation
Found in: 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
By Pushmeet Kohli,Anton Osokin,Stefanie Jegelka
Issue Date:June 2013
pp. 1971-1978
We discuss a model for image segmentation that is able to overcome the short-boundary bias observed in standard pairwise random field based approaches. To wit, we show that a random field with multi-layered hidden units can encode boundary preserving highe...
 
Spatial Inference Machines
Found in: 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
By Roman Shapovalov,Dmitry Vetrov,Pushmeet Kohli
Issue Date:June 2013
pp. 2985-2992
This paper addresses the problem of semantic segmentation of 3D point clouds. We extend the inference machines framework of Ross et al. by adding spatial factors that model mid-range and long-range dependencies inherent in the data. The new model is able t...
 
Learning an interactive segmentation system
Found in: Proceedings of the Seventh Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP '10)
By Carsten Rother, Christoph Rhemann, Hannes Nickisch, Pushmeet Kohli
Issue Date:December 2010
pp. 274-281
Many successful applications of computer vision to image or video manipulation are interactive by nature. However, parameters of such systems are often trained neglecting the user. Traditionally, interactive systems have been treated in the same manner as ...
     
Unwrap mosaics: a new representation for video editing
Found in: ACM Transactions on Graphics (TOG)
By Alex Rav-Acha, Andrew Fitzgibbon, Carsten Rother, Pushmeet Kohli
Issue Date:August 2008
pp. 1-49
We introduce a new representation for video which facilitates a number of common editing tasks. The representation has some of the power of a full reconstruction of 3D surface models from video, but is designed to be easy to recover from a priori unseen an...
     
On partial optimality in multi-label MRFs
Found in: Proceedings of the 25th international conference on Machine learning (ICML '08)
By Alexander Shekhovtsov, Carsten Rother, Philip Torr, Pushmeet Kohli, Vladimir Kolmogorov
Issue Date:July 2008
pp. 480-487
We consider the problem of optimizing multilabel MRFs, which is in general NP-hard and ubiquitous in low-level computer vision. One approach for its solution is to formulate it as an integer linear programming and relax the integrality constraints. The app...
     
 1