A support vector machine designed to identify breasts at high risk using multi-probe generated REIS signals: a preliminary assessment

Gur, David; Lederman, Dror; Dhurjaty, Sreeram; Sumkin, Jules H.; Zuley, Margarita L.

doi:10.1117/12.844452

Cited by 12 publications

(23 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2 shows the six sets of EIS signal magnitude sweeps between the center probe and each of the six ''external'' probes, for a participant with a biopsy-confirmed cancer and a participant determined by mammography to be negative. As previously discussed 35 and also shown in the authors' own studies, 16,17,35,48 these figures demonstrate that in general negative cases exhibit a higher level of electrical impedance symmetry between the two breasts than positive cases.…”

Section: The Multi-probe Reis Systemsupporting

confidence: 55%

“…The definitions and the detailed computational methods of the resulting REIS features have been described elsewhere. 16,17 For example, feature 1 was defined as the absolute difference among the ranges of all six resonance frequencies generated by six pairs of probes that are applied between the left (L) and right (R) breasts F 1 ¼ ð jDf L À Df R jÞ:…”

Section: Feature Extraction and Selectionmentioning

confidence: 99%

“…The SVM has been proven successful in a large number of classification applications, and has also been developed and shown to be efficient for the task in hand. 17 In our study, a polynomial functionbased statistical model [s = (a * b + c) d ] was selected to build the classification hyper-plane of the SVM classifier. The third classifier is a GMM.…”

Section: Feature Extraction and Selectionmentioning

confidence: 99%

“…35 Based on an underlying assumption that if indeed impedance-based measurements provide useful risk information, then it would be most pronounced as changes in signal and phase at and near the resonance frequency, the authors have been developing and testing a unique resonance-frequency-based electrical impedance spectroscopy (REIS) system for breast cancer risk stratification. 16,17,37,48 This approach also takes advantage of the expected symmetry levels (or not) between corresponding REIS signals as ascertained from two contra-lateral breasts of the same woman. The initial feature set proposed for this study for developing classifiers includes 35 REIS-based defined features.…”

Section: Introductionmentioning

confidence: 99%

“…The initial feature set proposed for this study for developing classifiers includes 35 REIS-based defined features. 17 During the preliminary evaluation of the first 100 examinations using different classifiers and a Genetic Algorithm (GA)-based feature pruning method, 17 the authors achieved AUCs in the range of 0.80-0.83 using support vector machine (SVM), 5 artificial neural network (ANN), 4 and Gaussian mixture models (GMMs) that were estimated using a greedy learning algorithm. 41,42 In this study, an enlarged data set to re-optimize these three (ANN, SVM, and GMM) classifiers was used.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Improving Breast Cancer Risk Stratification Using Resonance-Frequency Electrical Impedance Spectroscopy Through Fusion of Multiple Classifiers

et al. 2010

Self Cite

View full text Add to dashboard Cite

This study aims to improve breast cancer risk stratification. A seven-probe resonance-frequency-based electrical impedance spectroscopy (REIS) system was designed, assembled, and utilized to establish a data set of examinations from 174 women. Three classifiers, including artificial neural network (ANN), support vector machine (SVM), and Gaussian mixture model (GMM), were independently developed to predict the likelihood of each woman to be recommended for biopsy. The performances of these classifiers were compared, and seven fusion methods for integrating these classifiers were investigated. The results showed that among the three classifiers, the ANN yielded the highest performance with an area under the curve (AUC) of 0.81 for the receiver operating characteristic (ROC), while SVM and GMM achieved AUCs of 0.80 and 0.78, respectively. Improvements of up to 3% were obtained using fusion of the three classifiers, with the largest improvement obtained using either a "minimum score" rule or a "weighted sum" rule. Comparing different combinations of two out of the three classifiers, the weighted sum rule provided the most robust and consistent results, with AUCs of 0.81, 0.83, and 0.82 for the different combinations of ANN and SVM, ANN and GMM, and SVM and GMM, respectively. Furthermore, at 90% specificity, the ANN, the weighted sum- and min rule-based classifiers, all detected 67% of the verified cancer cases as compared with 50, 50, and 60% detection of the high risk cases, respectively. The study demonstrated that REIS examinations provide relevant information for developing breast cancer risk stratification tools and that using fusion of several not-fully-correlated classifiers can improve classification performance.

show abstract

Section: The Multi-probe Reis Systemsupporting

confidence: 55%

Section: Feature Extraction and Selectionmentioning

confidence: 99%

Section: Feature Extraction and Selectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Improving Breast Cancer Risk Stratification Using Resonance-Frequency Electrical Impedance Spectroscopy Through Fusion of Multiple Classifiers

et al. 2010

Self Cite

View full text Add to dashboard Cite

show abstract

Brain Haemorrhage Detection Through SVM Classification of Electrical Impedance Tomography Measurements

McDermott

Dunne

O’Halloran

et al. 2019

Brain and Human Body Modeling

View full text Add to dashboard Cite

Pilot study: electrical impedance based tissue classification using support vector machine classifier

Grewal

Golnaraghi

2014

IET Science, Measurement & Technology

View full text Add to dashboard Cite

Tissue classification using computer aided diagnosis can help automated decision making to aid clinical diagnosis. Classification of breast tissue based on spectral features of impedance loci has frequently been done to classify malignant tissue with further requirement of more complex classification methodologies needed to improve the characterisation. In current study, tissue classification is done using in vivo electrical impedance data of 18 human subjects, from four quadrants of breast, palm, nail, arm, bicep and classified using algorithms involving machine learning methodologies, specifically support vector machines (SVMs) that are supervised learning models. They consist of learning algorithms based on the principal of structural risk minimisation. Two methodologies of SVM have been used in this study: with data binning and data pruning and without data binning and data pruning. Data binning and data pruning have improved the sensitivity of the SVM from 76.76 to 89.23%, but the specificity has decreased from 76.23 to 74.15%. This is a pilot study towards testing the reliability of the developed electrical impedance measuring system and developing a data mining-based decision making system into an electrical impedance spectroscopy system, to help users (physicians) with tissue classification leading to reliable objective decision making.

show abstract

A support vector machine designed to identify breasts at high risk using multi-probe generated REIS signals: a preliminary assessment

Cited by 12 publications

References 0 publications

Improving Breast Cancer Risk Stratification Using Resonance-Frequency Electrical Impedance Spectroscopy Through Fusion of Multiple Classifiers

Improving Breast Cancer Risk Stratification Using Resonance-Frequency Electrical Impedance Spectroscopy Through Fusion of Multiple Classifiers

Brain Haemorrhage Detection Through SVM Classification of Electrical Impedance Tomography Measurements

Pilot study: electrical impedance based tissue classification using support vector machine classifier

Contact Info

Product

Resources

About