Results 1  10
of
16
Markov Random Field Modeling, Inference & Learning in Computer Vision & Image Understanding: A Survey
, 2013
"... ..."
Minimizing Sparse HighOrder Energies by Submodular Vertexcover
, 2012
"... Inference in highorder graphical models has become important in recent years. Several approaches are based, for example, on generalized messagepassing, or on transformation to a pairwise model with extra ‘auxiliary ’ variables. We focus on a special case where a much more efficient transformation ..."
Abstract

Cited by 5 (0 self)
 Add to MetaCart
Inference in highorder graphical models has become important in recent years. Several approaches are based, for example, on generalized messagepassing, or on transformation to a pairwise model with extra ‘auxiliary ’ variables. We focus on a special case where a much more efficient transformation is possible. Instead of adding variables, we transform the original problem into a comparatively small instance of submodular vertexcover. These vertexcover instances can then be attacked by existing algorithms (e.g. belief propagation, QPBO), where they often run 4–15 times faster and find better solutions than when applied to the original problem. We evaluate our approach on synthetic data, then we show applications within a fast hierarchical clustering and modelfitting framework.
Optimal Decisions from Probabilistic Models: the IntersectionoverUnion Case
"... A probabilistic model allows us to reason about the world and make statistically optimal decisions using Bayesian decision theory. However, in practice the intractability of the decision problem forces us to adopt simplistic loss functions such as the 0/1 loss or Hamming loss and as result we make ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
(Show Context)
A probabilistic model allows us to reason about the world and make statistically optimal decisions using Bayesian decision theory. However, in practice the intractability of the decision problem forces us to adopt simplistic loss functions such as the 0/1 loss or Hamming loss and as result we make poor decisions through MAP estimates or through loworder marginal statistics. In this work we investigate optimal decision making for more realistic loss functions. Specifically we consider the popular intersectionoverunion (IoU) score used in image segmentation benchmarks and show that it results in a hard combinatorial decision problem. To make this problem tractable we propose a statistical approximation to the objective function, as well as an approximate algorithm based on parametric linear programming. We apply the algorithm on three benchmark datasets and obtain improved intersectionoverunion scores compared to maximumposteriormarginal decisions. Our work points out the difficulties of using realistic loss functions with probabilistic computer vision models. 1.
Learning to Segment Neurons with nonlocal Quality Measures
"... Segmentation schemes such as hierarchical region merging or correllation clustering rely on edge weights between adjacent (super)voxels. The quality of these edge weights directly affects the quality of the resulting segmentations. Unstructured learning methods seek to minimize the classification ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
(Show Context)
Segmentation schemes such as hierarchical region merging or correllation clustering rely on edge weights between adjacent (super)voxels. The quality of these edge weights directly affects the quality of the resulting segmentations. Unstructured learning methods seek to minimize the classification error on individual edges. This ignores that a few local mistakes (tiny boundary gaps) can cause catastrophic global segmentation errors. Boundary evidence learning should therefore optimize structured quality criteria such as Rand Error or Variation of Information. We present the first structured learning scheme using a structured loss function; and we introduce a new hierarchical scheme that allows to approximately solve the NP hard prediction problem even for huge volume images. The value of these contributions is demonstrated on two challenging neural circuit reconstruction problems in serial sectioning electron microscopic images with billions of voxels. Our contributions lead to a partitioning quality that improves over the current state of the art.
Learning efficient random maximum aposteriori predictors with nondecomposable loss functions
 Advances in Neural Information Processing Systems
"... In this work we develop efficient methods for learning random MAP predictors for structured label problems. In particular, we construct posterior distributions over perturbations that can be adjusted via stochastic gradient methods. We show that every smooth posterior distribution would suffice to d ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
(Show Context)
In this work we develop efficient methods for learning random MAP predictors for structured label problems. In particular, we construct posterior distributions over perturbations that can be adjusted via stochastic gradient methods. We show that every smooth posterior distribution would suffice to define a smooth PACBayesian risk bound suitable for gradient methods. In addition, we relate the posterior distributions to computational properties of the MAP predictors. We suggest multiplicative posteriors to learn supermodular potential functions that accompany specialized MAP predictors such as graphcuts. We also describe labelaugmented posterior models that can use efficient MAP approximations, such as those arising from linear program relaxations. 1
High Order Regularization for SemiSupervised Learning of Structured Output Problems
"... Semisupervised learning, which uses unlabeled data to help learn a discriminative model, is especially important for structured output problems, as considerably more effort is needed to label its multidimensional outputs versus standard single output problems. We propose a new maxmargin framewor ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
(Show Context)
Semisupervised learning, which uses unlabeled data to help learn a discriminative model, is especially important for structured output problems, as considerably more effort is needed to label its multidimensional outputs versus standard single output problems. We propose a new maxmargin framework for semisupervised structured output learning, that allows the use of powerful discrete optimization algorithms and high order regularizers defined directly on model predictions for the unlabeled examples. We show that our framework is closely related to Posterior Regularization, and the two frameworks optimize special cases of the same objective. The new framework is instantiated on two image segmentation tasks, using both a graph regularizer and a cardinality regularizer. Experiments also demonstrate that this framework can utilize unlabeled data from a different source than the labeled data to significantly improve performance while saving labeling effort. 1.
Learning Structured Models with the AUC Loss and Its Generalizations
"... Many problems involve the prediction of multiple, possibly dependent labels. The structured output prediction framework builds predictors that take these dependencies into account and use them to improve accuracy. In many such tasks, performance is evaluated by the Area Under the ROC Curve (AUC). ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Many problems involve the prediction of multiple, possibly dependent labels. The structured output prediction framework builds predictors that take these dependencies into account and use them to improve accuracy. In many such tasks, performance is evaluated by the Area Under the ROC Curve (AUC). While a framework for optimizing the AUC loss for unstructured models exists, it does not naturally extend to structured models. In this work, we propose a representation and learning formulation for optimizing structured models over the AUC loss, show how our approach generalizes the unstructured case, and provide algorithms for solving the resulting inference and learning problems. We also explore several new variants of the AUC measure which naturally arise from our formulation. Finally, we empirically show the utility of our approach in several domains. 1
Learning with target prior
 In F
, 2012
"... In the conventional approaches for supervised parametric learning, relations between data and target variables are provided through training sets consisting of pairs of corresponded data and target variables. In this work, we describe a new learning scheme for parametric learning, in which the targ ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
(Show Context)
In the conventional approaches for supervised parametric learning, relations between data and target variables are provided through training sets consisting of pairs of corresponded data and target variables. In this work, we describe a new learning scheme for parametric learning, in which the target variables y can be modeled with a prior model p(y) and the relations between data and target variables are estimated with p(y) and a set of uncorresponded data X in training. We term this method as learning with target priors (LTP). Specifically, LTP learning seeks parameter θ that maximizes the log likelihood of fθ(X) on a uncorresponded training set with regards to p(y). Compared to the conventional (semi)supervised learning approach, LTP can make efficient use of prior knowledge of the target variables in the form of probabilistic distributions, and thus removes/reduces the reliance on training data in learning. Compared to the Bayesian approach, the learned parametric regressor in LTP can be more efficiently implemented and deployed in tasks where running efficiency is critical. We demonstrate the effectiveness of the proposed approach on parametric regression tasks for BCI signal decoding and pose estimation from video. 1
Instance Segmentation of Indoor Scenes using a Coverage Loss
"... Abstract. A major limitation of existing models for semantic segmentation is the inability to identify individual instances of the same class: when labeling pixels with only semantic classes, a set of pixels with the same label could represent a single object or ten. In this work, we introduce a m ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Abstract. A major limitation of existing models for semantic segmentation is the inability to identify individual instances of the same class: when labeling pixels with only semantic classes, a set of pixels with the same label could represent a single object or ten. In this work, we introduce a model to perform both semantic and instance segmentation simultaneously. We introduce a new higherorder loss function that directly minimizes the coverage metric and evaluate a variety of region features, including those from a convolutional network. We apply our model to the NYU Depth V2 dataset, obtaining state of the art results.
Perceptually Inspired Layoutaware Losses for Image Segmentation
"... Abstract. Interactive image segmentation is an important computer vision problem that has numerous real world applications. Models for image segmentation are generally trained to minimize the Hamming error in pixel labeling. The Hamming loss does not ensure that the topology/structure of the object ..."
Abstract
 Add to MetaCart
(Show Context)
Abstract. Interactive image segmentation is an important computer vision problem that has numerous real world applications. Models for image segmentation are generally trained to minimize the Hamming error in pixel labeling. The Hamming loss does not ensure that the topology/structure of the object being segmented is preserved and therefore is not a strong indicator of the quality of the segmentation as perceived by users. However, it is still ubiquitously used for training models because it decomposes over pixels and thus enables efficient learning. In this paper, we propose the use of a novel family of higherorder loss functions that encourage segmentations whose layout is similar to the groundtruth segmentation. Unlike the Hamming loss, these loss functions do not decompose over pixels and therefore cannot be directly used for lossaugmented inference. We show how our loss functions can be transformed to allow efficient learning and demonstrate the effectiveness of our method on a challenging segmentation dataset and validate the results using a user study. Our experimental results reveal that training with our layoutaware loss functions results in better segmentations that are preferred by users over segmentations obtained using conventional loss functions.