A fractal approach to the segmentation of microcalcifications in digital mammograms

Abstract

This paper presents a computerized method for the automated segmentation of individual microcalcifications in a region of interest (ROI) known to contain a cluster in digital mammograms. Mammographic parenchyma caj be accurately modeled with the fractal approach, but not areas with microcalcifications. The digitized image is divided into 16 x 16-pixel overlapping windows and those accurately modeled by the fractal model are eliminated. The next steps include local thresholding of the ROIs using an iterative method, the elimination of some of the artifacts and identification of the clustered microcalcifications using a clustering algorithm. The evaluation was performed on 81 simulated clusters superimposed on normal mammographic backgrounds and on a representative database of 408 real mammograms. Microcalcification locations were identified by two radiologists independently. These locations were compared to those found by the computer algorithm. An average of 59% of the simulated microcalcifications and 69% of the microcalcifications common to both radiologists were detected. The algorithm described provides a fully automated method for the segmentation of individual microcalcifications in an area of the mammogram known to contain a cluster.

An automatic microcalcification detection system based on a hybrid neural network classifier

“…Concerning image segmentation and specification of regions of interest (ROIs), several methods have been proposed such as classical image filtering and local thresholding [9,12,39,45], techniques based on mathematical morphology [13,60], stochastic fractal models [25,26], wavelet analysis [3,7,22,23,46,52,56,57] and multiscale analysis based on a specialized Gaussian and Peitgen [32]. Furthermore, various classification methodologies have been reported for the characterization of ROI such as, rule-based systems [9,12], fuzzy logic systems [11], statistical methods based on Markov random fields [20] and support vector machines [3].…”

Fractal analyses of osseous healing using Tuned Aperture Computed Tomography images

“…We believe that this study establishes the potential of TACT to quantify healing in osseous defects, thus providing alternatives to QCT and other advanced imaging methods [9,10,11,12,13,14,15,18,19,20,21]. The TACT technique does not require expensive equipment and can be easily employed in routine clinical and laboratory settings without having to standardize or maintain rigid projection geometry.…”

“…The higher the fractal dimension, the more the morphological complexity at the ultra-structural level [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]. Analyses of fractal dimension has been correlated with the strength of bone in previous studies [5,6,7,14,15,16,18], and therefore we decided to evaluate the fractal dimension of healing sites on rabbit mandibles which were subject to destructive bony lesion induction and various treatment modalities, using tuned aperture computed tomography (TACT; OrthoTACT, Instrumentarium Imaging, Helsinki, Finland) images.…”

New techniques for imaging, digitization and analysis of three-dimensional neural morphology on multiple scales

“…In particular, analysis of the spatial complexity of these structures has been performed either on 2D projections, or using skeletonized medial axes without accounting for the mass distribution. In recent work, 2D projections of tumor surfaces (Lefebvre et al, 1995;Pohlman et al, 1996; were segmented, then analyzed using planar box-counting, box-gliding or similar methods. Inferences of fractal dimensions and growth mechanisms based on such 2D data are not accurate representations of the true 3D spatial complexity, however Witten and Sander, 1983).…”