A Trained Spin-Glass Model for Grouping of Image Primitives

Image Sciences Institute, University Medical Center Utrecht, Heidelberglaan 100, E01.335, 3584 CX Utrecht, The Netherlands.
IEEE Transactions on Pattern Analysis and Machine Intelligence (Impact Factor: 5.78). 08/2005; 27(7):1172-82. DOI: 10.1109/TPAMI.2005.131
Source: PubMed


A method is presented that uses grouping to improve local classification of image primitives. The grouping process is based upon a spin-glass system, where the image primitives are treated as possessing a spin. The system is subject to an energy functional consisting of a local and a bilocal part, allowing interaction between the image primitives. Instead of defining the state of lowest energy as the grouping result, the mean state of the system is taken. In this way, instabilities caused by multiple minima in the energy are being avoided. The means of the spins are taken as the a posteriori probabilities for the grouping result. In the paper, it is shown how the energy functional can be learned from example data. The energy functional is defined in such a way that, in case of no interactions between the elements, the means of the spins equal the a priori local probabilities. The grouping process enables the fusion of the a priori local and bilocal probabilities into the a posteriori probabilities. The method is illustrated both on grouping of line elements in synthetic images and on vessel detection in retinal fundus images.

Download full-text


Available from: Joes Staal, Oct 10, 2015
16 Reads
  • Source
    • "A Metropolis algorithm is applied to find the expected values of the state variables of the system, that is, the probabilities for the labels of the primitives. For a detailed description of this classifier we refer to Staal et al. (2005). For the local features geometrical information of the primitives and intensity based information is extracted. "
    [Show abstract] [Hide abstract]
    ABSTRACT: A system for automatic segmentation and labeling of the complete rib cage in chest CT scans is presented. The method uses a general framework for automatic detection, recognition and segmentation of objects in three-dimensional medical images. The framework consists of five stages: (1) detection of relevant image structures, (2) construction of image primitives, (3) classification of the primitives, (4) grouping and recognition of classified primitives and (5) full segmentation based on the obtained groups. For this application, first 1D ridges are extracted in 3D data. Then, primitives in the form of line elements are constructed from the ridge voxels. Next a classifier is trained to classify the primitives in foreground (ribs) and background. In the grouping stage centerlines are formed from the foreground primitives and rib numbers are assigned to the centerlines. In the final segmentation stage, the centerlines act as initialization for a seeded region growing algorithm. The method is tested on 20 CT-scans. Of the primitives, 97.5% is classified correctly (sensitivity is 96.8%, specificity is 97.8%). After grouping, 98.4% of the ribs are recognized. The final segmentation is qualitatively evaluated and is very accurate for over 80% of all ribs, with slight errors otherwise.
    Medical Image Analysis 03/2007; 11(1):35-46. DOI:10.1016/ · 3.65 Impact Factor
  • Source
    • "Several articles have introduced techniques for vessel or ridge extraction based on the eigen-decomposition of the Hessian computed at each image pixel [38], [42], [49], [56]–[60]. We choose two specific measures here for our analysis, [42] and [49], which have been applied in a number of papers [38]–[40], [44], [45], [55]. They both start from the definition of the scale-space representation L : R 2 × R + → R L(x, y; t) = g(x, y; t) * f (x, y), (1) where g(·; t) is a Gaussian function with variance t, f is an image, (x, y) is a pixel location, and * represents the convolution operation [59]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Motivated by the goals of improving detection of low-contrast and narrow vessels and eliminating false detections at nonvascular structures, a new technique is presented for extracting vessels in retinal images. The core of the technique is a new likelihood ratio test that combines matched-filter responses, confidence measures and vessel boundary measures. Matched filter responses are derived in scale-space to extract vessels of widely varying widths. A vessel confidence measure is defined as a projection of a vector formed from a normalized pixel neighborhood onto a normalized ideal vessel profile. Vessel boundary measures and associated confidences are computed at potential vessel boundaries. Combined, these responses form a six-dimensional measurement vector at each pixel. A training technique is used to develop a mapping of this vector to a likelihood ratio that measures the "vesselness" at each pixel. Results comparing this vesselness measure to matched filters alone and to measures based on the Hessian of intensities show substantial improvements, both qualitatively and quantitatively. The Hessian can be used in place of the matched filter to obtain similar but less-substantial improvements or to steer the matched filter by preselecting kernel orientations. Finally, the new vesselness likelihood ratio is embedded into a vessel tracing framework, resulting in an efficient and effective vessel centerline extraction algorithm.
    IEEE Transactions on Medical Imaging 01/2007; 25(12):1531-46. DOI:10.1109/TMI.2006.884190 · 3.39 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We present an effective algorithm for automatic tracing of vasculature structures and vascular landmark extraction of bifurcations and ending points. In this paper we deal with vascular patterns from digital images for personal identification. Vessel tracing algorithms are of interest in a variety of biometric and medical application such as personal identification, biometrics, and ophthalmic disorders like vessel change detection. However eye surface vasculature tracing has many problems which are subject to improper illumination, glare, fade-out, shadow and artifacts arising from reflection, refraction, and dispersion. The proposed algorithm on vascular tracing employs multi-stage processing of ten-layers as followings: Image Acquisition, Image Enhancement by gray scale retinal image enhancement, reducing background artifact and illuminations and removing interlacing minute characteristics of vessels, Vascular Structure Extraction by connecting broken vessels, extracting vascular structure using eight directional information, and extracting retinal vascular structure. Vascular Landmark Extraction of bifurcations and ending points. The results of automatic retinal vessel extraction using five different thresholds applied 34 eye images are presented. The results of vasculature tracing algorithm shows that the suggested algorithm can obtain not only robust and accurate vessel tracing but also vascular landmarks according to thresholds.
Show more