Gautam S. Muralidhar

Blind Image Quality Assessment Without Human Training Using Latent Quality Factors

We propose a highly unsupervised, training free, no reference image quality assessment (IQA) model that is based on the hypothesis that distorted images have certain latent characteristics that differ from those of “natural” or “pristine” images. These latent characteristics are uncovered by applying a “topic model” to visual words extracted from an assortment of pristine and distorted images. For the latent characteristics to be discriminatory between pristine and distorted images, the choice of the visual words is important. We extract quality-aware visual words that are based on natural scene statistic features [1]. We show that the similarity between the probability of occurrence of the different topics in an unseen image and the distribution of latent topics averaged over a large number of pristine natural images yields a quality measure. This measure correlates well with human difference mean opinion scores on the LIVE IQA database [2].

Snakules: A Model-Based Active Contour Algorithm for the Annotation of Spicules on Mammography

We have developed a novel, model-based active contour algorithm, termed “snakules”, for the annotation of spicules on mammography. At each suspect spiculated mass location that has been identified by either a radiologist or a computer-aided detection (CADe) algorithm, we deploy snakules that are converging open-ended active contours also known as snakes. The set of convergent snakules have the ability to deform, grow and adapt to the true spicules in the image, by an attractive process of curve evolution and motion that optimizes the local matching energy. Starting from a natural set of automatically detected candidate points, snakules are deployed in the region around a suspect spiculated mass location. Statistics of prior physical measurements of spiculated masses on mammography are used in the process of detecting the set of candidate points. Observer studies with experienced radiologists to evaluate the performance of snakules demonstrate the potential of the algorithm as an image analysis technique to improve the specificity of CADe algorithms and as a CADe prompting tool.

Model-driven, probabilistic level set based segmentation of magnetic resonance images of the brain

Accurate segmentation of magnetic resonance (MR) images of the brain to differentiate features such as soft tissue, tumor, edema and necrosis is critical for both diagnosis and treatment purposes. Region-based formulations of geometric active contour models are popular choices for segmentation of MR and other medical images. Most of the traditional region-based formulations model local region intensity by assuming a piecewise constant approximation. However, the piecewise constant approximation rarely holds true for medical images such as MR images due to the presence of noise and bias field, which invariably results in a poor segmentation of the image. To overcome this problem, we have developed a probabilistic region-based active contour model for automatic segmentation of MR images of the brain. In our approach, a mixture of Gaussian distributions is used to accurately model the arbitrarily shaped local region intensity distribution. Prior spatial information derived from probabilistic atlases is also integrated into the level set evolution framework for guiding the segmentation process. Our experiments with a series of publicly available brain MR images show that the proposed active contour model gives stable and accurate segmentation results when compared to the traditional region based formulations.

A Steerable, Multiscale Singularity Index

We propose a new steerable, multiscale ratio index for detecting impulse singularities in signals of arbitrary dimensionality. For example, it responds strongly to curvilinear masses (ridges) in images, but minimally to step discontinuities. The ratio index employs directional derivatives of gaussians, making it naturally steerable and scalable. Experiments on real images demonstrate the efficacy of the index for detecting multiscale curvilinear structures. A software version of the index can be downloaded from: http://live.ece.utexas.edu/research/SingularityIndex/SingularityIndex.zip.

Computer-aided diagnosis in breast magnetic resonance imaging.

In this paper, we review the role played by breast magnetic resonance imaging in the detection and diagnosis of breast cancer. This is followed by a discussion of clinical decision support systems in medicine and their contributions in breast magnetic resonance imaging interpretation. We conclude by discussing the future of computer-aided diagnosis in breast magnetic resonance imaging.

Snakules for automatic classification of candidate spiculated mass locations on mammography

In this paper, we describe a novel approach for the automatic classification of candidate spiculated mass locations on mammography. Our approach is based on “Snakules” — an evidence-based active contour algorithm that we have recently developed for the annotation of spicules on mammography. We use snakules to extract features characteristic of spicules and spiculated masses, and use these features to classify whether a region of a mammogram contains a spiculated mass or not. The results from our initial classification experiment demonstrate the strong potential of snakules as an image analysis technique to extract features specific to spicules and spiculated masses, which can subsequently be used to distinguish true spiculated mass locations from non-lesion locations on a mammogram and improve the specificity of computer-aided detection (CADe) algorithms.

Stereoscopic interpretation of low-dose breast tomosynthesis projection images

The purpose of this study was to evaluate stereoscopic perception of low-dose breast tomosynthesis projection images. In this Institutional Review Board exempt study, craniocaudal breast tomosynthesis cases (N = 47), consisting of 23 biopsy-proven malignant mass cases and 24 normal cases, were retrospectively reviewed. A stereoscopic pair comprised of two projection images that were ±4° apart from the zero angle projection was displayed on a Planar PL2010M stereoscopic display (Planar Systems, Inc., Beaverton, OR, USA). An experienced breast imager verified the truth for each case stereoscopically. A two-phase blinded observer study was conducted. In the first phase, two experienced breast imagers rated their ability to perceive 3D information using a scale of 1–3 and described the most suspicious lesion using the BI-RADS® descriptors. In the second phase, four experienced breast imagers were asked to make a binary decision on whether they saw a mass for which they would initiate a diagnostic workup or not and also report the location of the mass and provide a confidence score in the range of 0–100. The sensitivity and the specificity of the lesion detection task were evaluated. The results from our study suggest that radiologists who can perceive stereo can reliably interpret breast tomosynthesis projection images using stereoscopic viewing.

Noise Analysis of a New Singularity Index

We analyze the noise sensitivity of a new singularity index that was designed to detect impulse singularities in signals of arbitrary dimensionality while rejecting step-like singularities see Muralidhar, IEEE Signal Process. Lett., vol. 20, no. 1, pp. 7-10, 2013 and Muralidhar , Proc. IEEE Int. Conf. Image Process., 2012. For example, the index responds strongly to curvilinear masses (ridges) in images, while weakly to jump discontinuities (edges). We analyze the detection power of the index in the presence of noise. Our analysis is geared towards answering the following questions: a) in the presence of noise only, what is the probability of falsely detecting an impulse given a threshold; b) given an impulse submerged in noise, what is the probability of detecting it given a threshold; and c) since the index is designed to be edge suppressing, what is the probability of incorrectly detecting an edge submerged in noise given a threshold. We compare the detection power of the index with that of a nominal impulse detector, the second derivative operator. Simulations and example applications in 1-D and 2-D reveal the efficacy of the new singularity index for correctly detecting impulses submerged in noise, while suppressing edges. A software version of the 2-D singularity index can be downloaded from: http://live.ece.utexas.edu/research/SingularityIndex/SingularityIndexCode.zip.

A shape constrained parametric active contour model for breast contour detection

Quantitative measures of breast morphology can help a breast cancer survivor to understand outcomes of reconstructive surgeries. One bottleneck of quantifying breast morphology is that there are only a few reliable automation algorithms for detecting the breast contour. This study proposes a novel approach for detecting the breast contour, which is based on a parametric active contour model. In addition to employing the traditional parametric active contour model, the proposed approach enforces a mathematical shape constraint based on the catenary curve, which has been previously shown to capture the overall shape of the breast contour reliably [1]. The mathematical shape constraint regulates the evolution of the active contour and helps the contour evolve towards the breast, while minimizing the undesired effects of other structures such as, the nipple/areola and scars. The efficacy of the proposed approach was evaluated on anterior posterior photographs of women who underwent or were scheduled for breast reconstruction surgery including autologous tissue reconstruction. The proposed algorithm shows promising results for detecting the breast contour.

A new singularity index

We propose a new ratio index for the detection of impulse-like singularities in signals of arbitrary dimensionality. We show that the new singularity index responds strongly to singularities that are like impulses or smoothed impulses in cross section. For example, it responds strongly to curvilinear masses (ridges) in images, while responding minimally to edge-like singularities. The ratio index employs directional derivatives of gaussians, which makes the index naturally scalable.