Pedro F. Felzenszwalb

Gabby Litterio, Juan-David Lizarazo-Ferro, Pedro Felzenszwalb et al.

We consider the limits of super-resolution using imaging constraints. Due to various theoretical and practical limitations, reconstruction-based methods have been largely restricted to small increases in resolution. In addition, motion-blur is usually seen as a nuisance that impedes super-resolution. We show that by using high-precision motion information, sparse image priors, and convex optimization, it is possible to increase resolution by large factors. A key operation in super-resolution is deconvolution with a box. In general, convolution with a box is not invertible. However, we obtain perfect reconstructions of sparse signals using convex optimization. We also show that motion blur can be helpful for super-resolution. We demonstrate that using pseudo-random motion it is possible to reconstruct a high-resolution target using a single low-resolution image. We present numerical experiments with simulated data and results with real data captured by a camera mounted on a computer controlled stage.

Anna Grim, Pedro Felzenszwalb

We present new message passing algorithms for performing inference with graphical models. Our methods are designed for the most difficult inference problems where loopy belief propagation and other heuristics fail to converge. Belief propagation is guaranteed to converge when the underlying graphical model is acyclic, but can fail to converge and is sensitive to initialization when the underlying graph has complex topology. This paper describes modifications to the standard belief propagation algorithms that lead to methods that converge to unique solutions on graphical models with arbitrary topology and potential functions.

Jeova F. S. Rocha Neto, Pedro Felzenszwalb, Marilyn Vazquez

Image segmentation algorithms often depend on appearance models that characterize the distribution of pixel values in different image regions. We describe a new approach for estimating appearance models directly from an image, without explicit consideration of the pixels that make up each region. Our approach is based on novel algebraic expressions that relate local image statistics to the appearance of spatially coherent regions. We describe two algorithms that can use the aforementioned algebraic expressions to estimate appearance models directly from an image. The first algorithm solves a system of linear and quadratic equations using a least squares formulation. The second algorithm is a spectral method based on an eigenvector computation. We present experimental results that demonstrate the proposed methods work well in practice and lead to effective image segmentation algorithms.

Pedro Felzenszwalb, Caroline Klivans, Alice Paul

We introduce a novel method for clustering using a semidefinite programming (SDP) relaxation of the Max k-Cut problem. The approach is based on a new methodology for rounding the solution of an SDP relaxation using iterated linear optimization. We show the vertices of the Max k-Cut relaxation correspond to partitions of the data into at most k sets. We also show the vertices are attractive fixed points of iterated linear optimization. Each step of this iterative process solves a relaxation of the closest vertex problem and leads to a new clustering problem where the underlying clusters are more clearly defined. Our experiments show that using fixed point iteration for rounding the Max k-Cut SDP relaxation leads to significantly better results when compared to randomized rounding.

Jeova F. S. Rocha Neto, Pedro F. Felzenszwalb

We introduce a new spectral method for image segmentation that incorporates long range relationships for global appearance modeling. The approach combines two different graphs, one is a sparse graph that captures spatial relationships between nearby pixels and another is a dense graph that captures pairwise similarity between all pairs of pixels. We extend the spectral method for Normalized Cuts to this setting by combining the transition matrices of Markov chains associated with each graph. We also derive an efficient method for sparsifying the dense graph of appearance relationships. This leads to a practical algorithm for segmenting high-resolution images. The resulting method can segment challenging images without any filtering or pre-processing.

Pedro F. Felzenszwalb

We consider a problem that involves finding similar elements in a collection of sets. The problem is motivated by applications in machine learning and pattern recognition. We formulate the similar elements problem as an optimization and give an efficient approximation algorithm that finds a solution within a factor of 2 of the optimal. The similar elements problem is a special case of the metric labeling problem and we also give an efficient 2-approximation algorithm for the metric labeling problem on complete graphs.

Jeroen Chua, Pedro F. Felzenszwalb

We describe a general framework for probabilistic modeling of complex scenes and inference from ambiguous observations. The approach is motivated by applications in image analysis and is based on the use of priors defined by stochastic grammars. We define a class of grammars that capture relationships between the objects in a scene and provide important contextual cues for statistical inference. The distribution over scenes defined by a probabilistic scene grammar can be represented by a graphical model and this construction can be used for efficient inference with loopy belief propagation. We show experimental results with two different applications. One application involves the reconstruction of binary contour maps. Another application involves detecting and localizing faces in images. In both applications the same framework leads to robust inference algorithms that can effectively combine local information to reason about a scene.

Sobhan Naderi Parizi, Kun He, Reza Aghajani et al.

Non-convex optimization is ubiquitous in machine learning. Majorization-Minimization (MM) is a powerful iterative procedure for optimizing non-convex functions that works by optimizing a sequence of bounds on the function. In MM, the bound at each iteration is required to \emph{touch} the objective function at the optimizer of the previous bound. We show that this touching constraint is unnecessary and overly restrictive. We generalize MM by relaxing this constraint, and propose a new optimization framework, named Generalized Majorization-Minimization (G-MM), that is more flexible. For instance, G-MM can incorporate application-specific biases into the optimization procedure without changing the objective function. We derive G-MM algorithms for several latent variable models and show empirically that they consistently outperform their MM counterparts in optimizing non-convex objectives. In particular, G-MM algorithms appear to be less sensitive to initialization.

Pedro F. Felzenszwalb, Benar F. Svaiter

We define two algorithms for propagating information in classification problems with pairwise relationships. The algorithms are based on contraction maps and are related to non-linear diffusion and random walks on graphs. The approach is also related to message passing algorithms, including belief propagation and mean field methods. The algorithms we describe are guaranteed to converge on graphs with arbitrary topology. Moreover they always converge to a unique fixed point, independent of initialization. We prove that the fixed points of the algorithms under consideration define lower-bounds on the energy function and the max-marginals of a Markov random field. The theoretical results also illustrate a relationship between message passing algorithms and value iteration for an infinite horizon Markov decision process. We illustrate the practical application of the algorithms under study with numerical experiments in image restoration, stereo depth estimation and binary classification on a grid.

Sobhan Naderi Parizi, Andrea Vedaldi, Andrew Zisserman et al.

Part-based representations have been shown to be very useful for image classification. Learning part-based models is often viewed as a two-stage problem. First, a collection of informative parts is discovered, using heuristics that promote part distinctiveness and diversity, and then classifiers are trained on the vector of part responses. In this paper we unify the two stages and learn the image classifiers and a set of shared parts jointly. We generate an initial pool of parts by randomly sampling part candidates and selecting a good subset using L1/L2 regularization. All steps are driven "directly" by the same objective namely the classification loss on a training set. This lets us do away with engineered heuristics. We also introduce the notion of "negative parts", intended as parts that are negatively correlated with one or more classes. Negative parts are complementary to the parts discovered by other methods, which look only for positive correlations.

Pedro F. Felzenszwalb, John G. Oberlin

We describe a framework for defining high-order image models that can be used in a variety of applications. The approach involves modeling local patterns in a multiscale representation of an image. Local properties of a coarsened image reflect non-local properties of the original image. In the case of binary images local properties are defined by the binary patterns observed over small neighborhoods around each pixel. With the multiscale representation we capture the frequency of patterns observed at different scales of resolution. This framework leads to expressive priors that depend on a relatively small number of parameters. For inference and learning we use an MCMC method for block sampling with very large blocks. We evaluate the approach with two example applications. One involves contour detection. The other involves binary segmentation.

Pedro F. Felzenszwalb

We consider object detection using a generic model for natural shapes. A common approach for object recognition involves matching object models directly to images. Another approach involves building intermediate representations via a generic grouping processes. We argue that these two processes (model-based recognition and grouping) may use similar computational mechanisms. By defining a generic model for shapes we can use model-based techniques to implement a mid-level vision grouping process.