&lt;title&gt;Probabilistic approach to model extraction from training data&lt;/title&gt;

DeVore, Michael D.; O'Sullivan, Joseph A.; Anand, Sushil; Schmid, Natalia A.

doi:10.1117/12.438229

Cited by 8 publications

(9 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the limit, as the approximation intervals grow small and corresponding computation time grows large, the search results approach the true maximum of (7). DeVore, et al 3 show in a four-class MSTAR experiment that the probability of classification error decreases much more quickly as a function of bits transmitted from the database under this hierarchical approach than under an exhaustive search through azimuth with a fixed approximation, d m .…”

Section: Successively-refinable Parameter Encodingmentioning

confidence: 96%

“…The resulting performance-complexity curves provide a means of directly comparing the capabilities of alternate ATR algorithms without being overly sensitive to particular choices of implementation parameters. This paper builds upon that work by incorporating hierarchical target models as in DeVore, et al 3 rather than multiple, fixed-complexity representations. Moreover, models for several computation architectures are developed and used to predict processing rates for target images as a function of target model complexity.…”

Section: Introductionmentioning

confidence: 95%

“…This target model can be used to complete the SAR image model in (1) for any given pose, Θ, by applying the rotation and shift operations implied by Θ to the target model and projecting the result onto the radar slant plane corresponding to the given depression angle. 3 Since we are considering a fixed depression angle and only azimuthal target rotation, we can restrict our attention to models of targets in the slant plane. i (φ), is approximated as piecewise constant in viewing angle over N w intervals and its value over the kth interval, centered at φ k = 2πk/N w , is estimated from a subset of registered training images collected with azimuth angles close to φ k .…”

Section: Sar Model and Recognition Algorithmsmentioning

confidence: 99%

See 2 more Smart Citations

Relationships between computational system performance and recognition system performance

DeVore¹,

O'Sullivan²,

Chamberlain³

et al. 2001

SPIE Proceedings

Self Cite

View full text Add to dashboard Cite

The implementation of computational systems to perform intensive operations often involves balancing the performance specification, system throughput, and available system resources. For problems of automatic target recognition (ATR), these three quantities of interest are the probability of classification error, the rate at which regions of interest are processed, and the computational power of the underlying hardware. An understanding of the inter-relationships between these factors can be an aid in making informed choices while exploring competing design possibilities. To model these relationships we have combined characterizations of ATR performance, which yield probability of classification error as a function of target model complexity, with analytical models of computational performance, which yield throughput as a function of target model complexity. Together, these constitute a parametric curve that is parameterized by target model complexity for any given recognition problem and hardware implementation. We demonstrate this approach on the problem of ATR from synthetic aperture radar imagery using a subset of the publicly released MSTAR dataset. We use this approach to characterize the achievable classification rate as a function of required throughput for various hardware configurations.

show abstract

Section: Successively-refinable Parameter Encodingmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 95%

Section: Sar Model and Recognition Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Relationships between computational system performance and recognition system performance

DeVore¹,

O'Sullivan²,

Chamberlain³

et al. 2001

SPIE Proceedings

Self Cite

View full text Add to dashboard Cite

show abstract

“…9 These functions then define a target model for the object rather than a sensor model. The idea is that training images of objects are registered to a common orientation in the image plane before the parameter functions are estimated.…”

Section: Target-centered Geometrymentioning

confidence: 99%

<title>Conditionally gamma- and k-distribution models for SAR ATR</title>

DeVore¹,

O'Sullivan²,

Montagnino³

2002

Algorithms for Synthetic Aperture Radar Imagery IX

Self Cite

View full text Add to dashboard Cite

Model-based approaches to automatic target recognition (ATR) generally infer the class and pose of objects in imagery by exploiting theoretical models of the formed images. Recently, we have performed an evaluation of several statistical models for synthetic aperture radar (SAR) and have conducted experiments with ATR algorithms derived from these models. In particular, a one-parameter complex Gaussian model, classically used to model diffuse scattering, was shown to deliver higher recognition rates than a one-parameter quarter-power normal model on actual SAR data. However an extended, two-parameter quarter-power model was consistently a better fit to the data than a corresponding two-parameter Gaussian model.In this paper, we apply Rician, gamma, and K distribution models, which are two-parameter extensions of the complex Gaussian and quarter-power normal models, to ATR from SAR magnitude imagery. We consider maximum-likelihood estimation of unknown model parameters and apply the resulting training and testing algorithms to actual SAR data. We show that the K distribution model performs better than the Rician and gamma models for both large and small sample sizes. The one-parameter complex Gaussian model performed slightly better than the K model overall. For small sample sizes, this is likely due to the relative stability in estimating only one model parameter. For large sample sizes this is likely due to a lack of persistence in specular reflections over the large angular intervals required to obtain large samples.

show abstract

“…In evaluating a general approach to ATR, we seek to decouple the effects of these parameters in order to understand their relative importance. DeVore, et al [3] consider ATR performance as a function of resolution in pose space. In this section, we consider issues related to variation in segmentation.…”

Section: Performance Sensitivitymentioning

confidence: 99%

Target-Centered Models and Information-Theoretic Segmentation for Automatic Target Recognition

DeVore

O’Sullivan

2003

Radar Signal Processing and Its Applications

Self Cite

View full text Add to dashboard Cite

Abstract. We present an approach to automatic target recognition (ATR) from synthetic aperture radar (SAR) imagery which combines advantages of both model-based and template-based approaches. Prior observations are used to estimate the statistical properties of reflectance over regions in the training scene. These target-centered statistical models can then be used to estimate the statistical properties of sensor output for arbitrary pose. Twosided hypothesis tests which are maximally powerful at the most likely alternative are developed in a informationtheoretic framework to address target model segmentation and confuser rejection. Segmentation of target from clutter is performed in the target-centered coordinate system using all prior observations to produce a consistent segmentation over all poses. We present performance and computation complexity results as a function of segmentation threshold, confuser-rejection threshold, and operating conditions for publicly available SAR data.

show abstract

<title>Probabilistic approach to model extraction from training data</title>

Cited by 8 publications

References 9 publications

Relationships between computational system performance and recognition system performance

Relationships between computational system performance and recognition system performance

<title>Conditionally gamma- and k-distribution models for SAR ATR</title>

Target-Centered Models and Information-Theoretic Segmentation for Automatic Target Recognition

Contact Info

Product

Resources

About