Supervised learning of large perceptual organization: graph spectral partitioning and learning automata

Sarkar, Sudeep; Soundararajan, Padmanabhan

doi:10.1109/34.857006

Cited by 169 publications

(99 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To mitigate this bias, the graph cut cost can be normalized using the edge weights being cut and/or properties of the resulting regions. Although many cost functions have been proposed (e.g., [5], [11], [19], [25]), the most popular normalized cut formulation, referred to widely as N-Cuts, is due to Shi and Malik [21] and was the basis for the original superpixel algorithm of [18].…”

Section: Introductionmentioning

confidence: 99%

TurboPixels: Fast Superpixels Using Geometric Flows

Levinshtein

Stere

Kutulakos

et al. 2009

IEEE Trans. Pattern Anal. Mach. Intell.

1,056

676

View full text Add to dashboard Cite

Abstract-We describe a geometric-flow-based algorithm for computing a dense oversegmentation of an image, often referred to as superpixels. It produces segments that, on one hand, respect local image boundaries, while, on the other hand, limiting undersegmentation through a compactness constraint. It is very fast, with complexity that is approximately linear in image size, and can be applied to megapixel sized images with high superpixel densities in a matter of minutes. We show qualitative demonstrations of high-quality results on several complex images. The Berkeley database is used to quantitatively compare its performance to a number of oversegmentation algorithms, showing that it yields less undersegmentation than algorithms that lack a compactness constraint while offering a significant speedup over N-cuts, which does enforce compactness.

show abstract

Section: Introductionmentioning

confidence: 99%

TurboPixels: Fast Superpixels Using Geometric Flows

Levinshtein

Stere

Kutulakos

et al. 2009

IEEE Trans. Pattern Anal. Mach. Intell.

1,056

676

View full text Add to dashboard Cite

show abstract

“…In this section we provide an empirical demonstration of the process of construction and performance of the PDN using (i) a small vision modules (Canny edge detector) with 4 parameters, (ii) a coupling of two vision modules (Etemadi et al's grouper [32]) with 7 parameters, and (iii) a combination of three vision modules (graph spectral-based complex grouper [33,34]) with 21 parameters. We allow for 10 possible choices of each parameter.…”

Section: Resultsmentioning

confidence: 99%

“…The graph spectra-based grouping algorithm described in [33,34] is a good example of a complex vision subsystem that goes beyond the low level. The combination tries to form large groupings of constant curvature edge segments and consists of an edge detection module, a contour segmentation module, and a grouping module, which is based on the graph spectrum.…”

Section: The Vision Subsystems and Evaluation Measuresmentioning

confidence: 99%

“…The combination tries to form large groupings of constant curvature edge segments and consists of an edge detection module, a contour segmentation module, and a grouping module, which is based on the graph spectrum. We refer the reader to [33,34] for details about the algorithm. Here we are concerned just with its parameter space behavior over the 21 parameters that can be chosen.…”

Section: The Vision Subsystems and Evaluation Measuresmentioning

confidence: 99%

See 1 more Smart Citation

Modeling Parameter Space Behavior of Vision Systems Using Bayesian Networks

Sarkar

Chavali

2000

Computer Vision and Image Understanding

View full text Add to dashboard Cite

The performance of most vision systems (or subsystems) is significantly dependent on the choice of its various parameters or thresholds. The associated parameter search space is extremely large and nonsmooth; moreover, the optimal choices of the parameters are usually mutually dependent on each other. In this paper we offer a Bayesian network-based probabilistic formalism, which we call the parameter dependence networks (PDNs), to model, abstract, and analyze the parameter space behavior of vision systems. The various algorithm parameters are the nodes of the PDN and are associated with probabilistic beliefs about the optimality of their respective values. The links between the nodes capture the direct dependencies between them and are quantified by conditional belief functions. The PDN structure captures the interdependence among the parameters in a concise and explicit manner. We define information theoretic measures, based on these PDNs, to quantify the global parameter sensitivity and the strength of the interdependence of the parameters. These measures predict the general ease of parameter tuning and performance stability of the system. The PDNs can also be used to stochastically sample the parameter space, to select optimal parameter sets (e.g., in performance evaluation studies), and to choose parameters, given constraints on the choice of some parameters. We also offer a strategy based on stochastic learning automata to generate training data to automatically build these PDNs. The team of learning automata stochastically samples the parameter space in a nonuniform manner with more samples near optimal values. These nonuniform samples are used to infer both the dependency structure and the conditional probabilities of the PDN. We demonstrate the process of construction of the PDN for an isolated vision module with 4 parameters (an edge detector), a coupling of two vision modules with a total of 7 parameters (a small edge grouping module), and a combination of three vision modules with 21 parameters (a complex perceptual organization module).

show abstract

“…Outdoor images are then classified as city or landscape and are further divided into sunset, forest, and mountain classes. Statistical classification methods are applied to low level visual features to derive categories [18,6]. For example, The SemQuery system [20] categorizes images into different clusters based on their heterogeneous features.…”

Section: Introductionmentioning

confidence: 99%

An attention based similarity measure used to identify image clusters

Bamidele¹,

Stentiford²

2005

2nd European Workshop on the Integration of Knowledge, Semantics and Digital Media Technology (EWIMT 2005)

View full text Add to dashboard Cite

This paper outlines a new attention based similarity measure and describes an application to the problem of identifying image clusters in a 4 class problem. A diverse set of images was obtained using camera phones in 4 separate locations and classification performance was tested against the true location of the images. The approach promises to have application to the unsupervised extraction of unknown numbers of clusters in larger datasets.

show abstract

Supervised learning of large perceptual organization: graph spectral partitioning and learning automata

Cited by 169 publications

References 47 publications

TurboPixels: Fast Superpixels Using Geometric Flows

TurboPixels: Fast Superpixels Using Geometric Flows

Modeling Parameter Space Behavior of Vision Systems Using Bayesian Networks

An attention based similarity measure used to identify image clusters

Contact Info

Product

Resources

About