Segmenting mammographic microcalcifications using a semi-automatic procedure based on Otsu s method and morphological filters

A two‐stage method for microcalcification cluster segmentation in mammography by deformable models

“…AOM cluster values ranged from 0.59 to 0.64 for intra and interobserver agreement analyses, respectively, being similar to the results reported by other studies. 15,42 In case of amorphous MC clusters, a trend for decreased values of AOM cluster was observed, as compared to other morphology types of MC clusters, mainly attributed to inherent morphological characteristics of this specific morphology type, challenging segmentation.…”

“…Quantitative metrics have been utilized to assess accuracy of segmentation methods on real and simulated MCs in mammography. 15,16,18,42 In this study, in absence of ground truth with respect to actual individual MC boundaries, quantitative metrics are utilized to quantify inter and intraobserver agreements for the proposed segmentation method. Two radiologists (denoted as R1 and R2), experienced in mammography (6 yr for R1 and 14 yr for R2), were independently asked to segment MC particles within clusters.…”

“…Segmentation of individual MCs has been achieved by region growing, 6 grey-level thresholding on preprocessed regions of interest (ROIs), 5,7-12 morphologic operations, 3,[13][14][15] watershed algorithms, 16,17 Bayesian pixel classification combined with Markov random field models, 18 radial gradientbased methods, 17 and active contours. 19,20 Furthermore, the wavelet transform has been used for segmentation of MCs, due to its ability to spatially localize high-frequency components.…”

Assistive system based on nerve detection and needle navigation in ultrasound images for regional anesthesia

“…To deal with this problem, many of the methods proposed in the literature have used semi-automatic segmentation (Rabben et al, 2000;Mahdavi et al, 2011;Duarte et al, 2013;Ni et al, 2015), in which a human localizes a suitable region, and then segmentation algorithms delineate the selected region. This study was designed to automatically detect ROIs and delineate their boundaries.…”

Synthesis of Mammographic Images Based on the Fractional Brownian Motion

“…In the literature, several approaches were proposed to segment microcalcifications [15][16][17][18] such as active contours [16,19], curvelet moments [20], wavelet analysis [21][22][23], fractal analysis [24][25][26], multifractal analysis [27,28] and morphological filters [29][30][31][32] in order to reduce human subjectivity in diagnosis.…”