Enrique Corona

Digital stereo image analyzer for generating automated 3-D measures of optic disc deformation in glaucoma

The major limitations of precise evaluation of retinal structures in present clinical situations are the lack of standardization, the inherent subjectivity involved in the interpretation of retinal images, and intra- as well as interobserver variability. While evaluating optic disc deformation in glaucoma, these limitations could be overcome by using advanced digital image analysis techniques to generate precise metrics; from stereo optic disc image pairs. A digital stereovision system for visualizing the topography of the optic nerve head from stereo optic disc images is presented. We have developed an algorithm, combining power cepstrum and zero-mean-normalized cross correlation techniques, which extracts depth information using coarse-to-fine disparity between corresponding windows in a stereo pair. The gray level encoded sparse disparity matrix is subjected to a cubic B-spline operation to generate smooth representations of the optic cup/disc surfaces and new three-dimensional (3-D) metrics from isodisparity contours. Despite the challenges involved in 3-D surface recovery, the robustness of our algorithm in finding disparities within the constraints used has been validated using stereo pairs with known disparities. In a preliminary longitudinal study of glaucoma patients, a strong correlation is found between the computer-generated quantitative cup/disc volume metrics and manual metrics commonly used in a clinic. The computer generated new metrics, however, eliminate the subjective variability and greatly reduce the time and cost involved in manual metric generation in follow-up studies of glaucoma.

A Unified Model-Based Image Analysis Framework for Automated Detection of Precancerous Lesions in Digitized Uterine Cervix Images

A unified framework for a fully automated diagnostic system for cervical intraepithelial neoplasia (CIN) is proposed. CIN is a detectable and treatable precursor pathology of cancer of the uterine cervix. Algorithms based on mathematical morphology, and clustering based on Gaussian mixture modeling (GMM) in a joint color and geometric feature space, are used to segment macro regions. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion-rate) curve is proposed to assess the model order. This technique provides good starting points to infer the GMM parameters via the expectation-maximization (EM) algorithm, reducing the segmentation time and the chances of getting trapped in local optima. The classification of vascular abnormalities in CIN, such as mosaicism and punctations, is modeled as a texture classification problem, and a solution is attempted by characterizing the neighborhood gray-tone dependences and co-occurrence statistics of the textures. The model presented in this paper provides a sequential framework for translating digital images of the cervix into a complete diagnostic tool, with minimal human intervention. In its current form, the research presented in this work may be used to aid physicians to locate abnormalities due to CIN and assess the best areas for a biopsy.

Non-parametric Estimation of Mixture Model Order

Mixture models are among the most popular and effective techniques for image segmentation. While Gaussian Mixture Models (GMM) are a reasonable choice, the number of components is not easy to determine. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion- rate) curve is proposed for model order identification purposes. This curve is estimated via the popular K- means clustering algorithm. To achieve repeatability and efficiency, various centroid initialization and image down sampling methods are proposed and tested. This technique also provides good starting points for inferring the GMM parameters via the expectation-maximization (EM) algorithm, which effectively reduces the segmentation time and the chances of getting trapped in local optima.

Multilevel wavelet feature statistics for efficient retrieval, transmission, and display of medical images by hybrid encoding

Many common modalities of medical images acquire high-resolution and multispectral images, which are subsequently processed, visualized, and transmitted by subsampling. These subsampled images compromise resolution for processing ability, thus risking loss of significant diagnostic information. A hybrid multiresolution vector quantizer (HMVQ) has been developed exploiting the statistical characteristics of the features in a multiresolution wavelet-transformed domain. The global codebook generated by HMVQ, using a combination of multiresolution vector quantization and residual scalar encoding, retains edge information better and avoids significant blurring observed in reconstructed medical images by other well-known encoding schemes at low bit rates. Two specific image modalities, namely, X-ray radiographic and magnetic resonance imaging (MRI), have been considered as examples. The ability of HMVQ in reconstructing high-fidelity images at low bit rates makes it particularly desirable for medical image encoding and fast transmission of 3D medical images generated from multiview stereo pairs for visual communications.

Digital stereo-optic disc image analyzer for monitoring progression of glaucoma

This paper describes an automated 3-D surface recovery algorithm for consistent and quantitative evaluation of the deformation in the ONH (optic nerve head). Additional measures, such as the changes in the volume of the cup and the disc as an improvement to the traditional cup to disc ratios, can thus be developed for longitudinal follow-up study of a patient. We propose an automated computerized technique for stereo pair registration and surface visualization of the ONH. Power cepstrum and zero mean cross correlation are embedded in the registration and a 3-D surface recovery technique is proposed. Preprocessing, as well as an overall registration, is performed upon stereo pairs. Then a coarse to fine feature matching strategy is used to reduce the ambiguity in finding the conjugate pair of the same point within the constraints of the epipolar plane. A cubic B-spline interpolation smooths the representation of the ONH obtained, while superimposition of features such as blood vessels is added. Studies show high correlation between traditional cup/disc measures derived from manual segmentation by ophthalmologists and computer generated cup/disc volume ratio. Such longitudinal studies over a large population of glaucoma patients are currently in progress for validation of the surface recovery algorithm.

An information theoretic approach to automated medical image segmentation

Automated segmentation of medical images is a challenging problem. The number of segments in a medical image may be unknown a priori, due to the presence or absence of pathological anomalies. Some unsupervised learning techniques founded in information theory concepts may provide a solid approach to this problem’s solution. We have developed the Improved “Jump” Method (IJM), a technique that efficiently finds a suitable number of clusters representing different tissue characteristics in a medical image. IJM works by optimizes an objective function that quantifies the quality of particular cluster configurations. Recent developments involving interesting relationships between Spectral Clustering (SC) and kernel Principal Component Analysis (kPCA) are used to extend IJM to the non-linear domain. This novel SC approach maps the data to a new space where the points belonging to the same cluster are collinear if the parameters of a Radial Basis Function (RBF) kernel are adequately selected. After projecting these points onto the unit sphere, IJM measures the quality of different cluster configurations, yielding an algorithm that simultaneously selects the number of clusters, and the RBF kernel parameter. Validation of this method is sought via segmentation of MR brain images in a combination of all major modalities. Such labeled MRI datasets serve as benchmarks for any segmentation algorithm. The effectiveness of the nonlinear IJM is demonstrated in the segmentation of uterine cervix color images for early identification of cervical neoplasia, as an aid to cervical cancer diagnosis. Studies are in progress in segmentation and detection of multiple sclerosis lesions.

Information Theoretic Clustering for Medical Image Segmentation

Analysis of medical images involves robust computational approaches to various optimization problems prior to interpretation of the embedded pathological changes. Major computational efforts are essential in unsupervised learning of structures in various medical images. Some unsupervised learning techniques that take advantage of information theory concepts provide a different perspective on the solution of automated learning methods. This chapter will review a recent approach to clustering examined under an information theoretic framework that efficiently finds a suitable number of clusters representing different tissue characteristics in a medical image. The proposed clustering approach optimizes an objective function which quantifies the quality of particular cluster configurations. Recent developments involving interesting relationships between spectral clustering (SC) and kernel principal component analysis (kPCA) are used in this technique to include the nonlinear domain. In this novel SC approach, the data is mapped to a new space where the points belonging to the same cluster are collinear if the parameters of a radial basis function (RBF) kernel are adequately selected. The effectiveness of this nonlinear approach is demonstrated in the segmentation of uterine cervix color images for early identification of cervical neoplasia, as an aid to cervical cancer diagnosis. The limitations of this method in the segmentation of specific medical images such as brain images with multiple sclerosis lesions and a strategy to overcome them are discussed.

Random Access Region of Interest in Backward Coding of Wavelet Trees

Random access to high quality regions of interest (ROI) from compressed bit streams of large images is becoming a necessary feature in many applications, particularly in viewing important areas within larger images. We have developed a completely new multi-rate image subband coding scheme using backward coding of wavelet trees (BCWT), which is fast, memory-efficient and resolution-scalable, while offering much less complexity than many other codecs, including block-based ones, e.g. JPEG2000. Although the focus of this paper is the inclusion of random access ROI decoding capabilities in BCWT, the method can also be modified in order to accommodate conventional ROI schemes in which the selected region is decoded with higher fidelity than the rest of the image. Experimental results compare the original BCWT bitstream size against that of the resulting ROI.

Feature extraction and segmentation in medical images by statistical optimization and point operation approaches

Feature extraction is a critical preprocessing step, which influences the outcome of the entire process of developing significant metrics for medical image evaluation. The purpose of this paper is firstly to compare the effect of an optimized statistical feature extraction methodology to a well designed combination of point operations for feature extraction at the preprocessing stage of retinal images for developing useful diagnostic metrics for retinal diseases such as glaucoma and diabetic retinopathy. Segmentation of the extracted features allow us to investigate the effect of occlusion induced by these features on generating stereo disparity mapping and 3-D visualization of the optic cup/disc. Segmentation of blood vessels in the retina also has significant application in generating precise vessel diameter metrics in vascular diseases such as hypertension and diabetic retinopathy for monitoring progression of retinal diseases.

A fast algorithm for registration of individual frames and information recovery in fluorescein angiography video image analysis

A fast algorithm, based on power cepstrum and crosscorrelation, techniques, for registration of image frames in a video is presented. Since compensation for unwanted rotations between image frames is needed, the shift invariance property of Fourier spectrum is used to find this rotation. The registration and information recovery approaches involve three image processing steps: a feature extraction step using the difference of Gaussian filtered images combined with median filtering for removal of noise, to enhance significant fluorescent blood vessels combined with median filtering for removal of noise; compensation for rotational and translational shifts in the Fourier domain using the spectral properties, crosscorrelation, and power cepstrum; lastly, information in each frame is maximized in comparison to the subsequent frames. Such analysis allows better interpretation of the video content for diagnostic evaluation of retinopathies.