#### DMCA

## Mass classification in breast DCE-MR images using an artificial neural network trained via a bee colony optimization algorithm

### Citations

298 | Feature Selection for Classification - Dash, Liu - 1997 |

253 | An idea based on honey bee swarm for numerical optimization,” - Karaboga - 2005 |

207 | Texture analysis. In: - Tuceryan, AK - 1998 |

184 | Benchmarking Attribute Selection Techniques for Discrete Class Data Mining - Hall, Holmes |

173 | A powerful and efficient algorithm for numerical function optimization: Artificial bee colony (ABC) algorithm - Karaboga, Basturk - 2007 |

102 | Applied Logistic Regression, 2nd edn. - Hosmer, Lemeshow - 2000 |

94 | Textural features for image classification. - RM, Shanmugam, et al. - 1973 |

42 | An artificial bee colony (ABC) algorithm for numerical function optimization. In: - Basturk, Karaboga - 2006 |

30 | ROC analysis applied to the evaluation of medical imaging techniques. - JA - 1979 |

29 |
The current status of breast MR imaging Part I. Choice of technique, image interpretation, diagnostic accuracy, and transfer to clinical practice.
- Kuhl
- 2007
(Show Context)
Citation Context ...viour of honey bees and could be regarded as belonging to the category of intelligent optimization tools22. In the artificial bee colony (ABC) algorithm, the colony of artificial bees contains three groups of bees: employed bees, onlookers, and scouts. Each cycle of the search consists of three major steps: (1) placing the employed bees onto the food sources and then calculating their nectar amounts; (2) selecting the food sources by the onlookers after sharing the information of employed bees and determining the www.scienceasia.org 296 ScienceAsia 39 (2013) nectar amount of the food sources; (3) determining the scout bees and placing them onto the randomly determined food sources. In the ABC, a food source position represents a possible solution to the problem to be optimized and the nectar amount of a food source corresponds to the quality (fitness) of the associated solution23–25. The ABC algorithm performs a neighbourhood search combined with random search in a way that is indicative of the food foraging behaviour of swarms of honey bees. The algorithm has been successfully applied to different optimization problems including the training of neural networks for control chart patte... |

22 | Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes. - JV - 1996 |

21 | The Bees Algorithm—A novel tool for complex optimisation problems. In: - DT, Ghanbarzadeh, et al. - 2006 |

19 |
A fuzzy c-means (FCM)-based approach for computerized segmentation of breast lesions in dynamic contrast-enhanced MR images.
- Chen, ML, et al.
- 2006
(Show Context)
Citation Context ... the data that determine classifier performance. Determining a suitable classifier for a given problem is however more an art than a science. Classifier performance is a function of several factors including the statistical distribution of the training and testing data, the internal structure of the classifier and the inherent randomness in the training process. The classification performance can be assessed in terms of the sensitivity, specificity and accuracy of the system. Sensitivity (SN) is the proportion of actual positives which are correctly identified and it is mathematically defined (6) and specificity (SP) is the proportion of negatives which are correctly identified and is mathematically defined in (7): Sensitivity: SN = TP TP + FN (6) Specificity: SP = TN TN + FP (7) Accuracy: ACC = TP + TN TP + FP + FN + TN (8) Error in classification: E = 1−ACC, (9) where TP = true positive, TN = true negative, FP = false positive, and FN = false negative. It is obvious that the main objective of a classifier is to minimize the false positive and negative rates, similarly, to maximize the true negative and positive rates. The sensitivity, specificity, accuracy, and error of www.sciencea... |

13 | Texture analysis of contrast-enhanced MR images of the breast. - Gibbs, LW - 2003 |

12 |
Automatic identification and classification of characteristic kinetic curves of breast lesions on DCE-MRI.
- Chen, ML, et al.
- 2006
(Show Context)
Citation Context ...n art than a science. Classifier performance is a function of several factors including the statistical distribution of the training and testing data, the internal structure of the classifier and the inherent randomness in the training process. The classification performance can be assessed in terms of the sensitivity, specificity and accuracy of the system. Sensitivity (SN) is the proportion of actual positives which are correctly identified and it is mathematically defined (6) and specificity (SP) is the proportion of negatives which are correctly identified and is mathematically defined in (7): Sensitivity: SN = TP TP + FN (6) Specificity: SP = TN TN + FP (7) Accuracy: ACC = TP + TN TP + FP + FN + TN (8) Error in classification: E = 1−ACC, (9) where TP = true positive, TN = true negative, FP = false positive, and FN = false negative. It is obvious that the main objective of a classifier is to minimize the false positive and negative rates, similarly, to maximize the true negative and positive rates. The sensitivity, specificity, accuracy, and error of www.scienceasia.org ScienceAsia 39 (2013) 301 the classification technique were evaluated through quantitative measures derived thro... |

12 | A review of evidence of health benefits from artificial neural networks in medical intervention. - PJG - 2002 |

9 | Optimising neural networks for identification of wood defects using the Bees Algorithm. In: - DT, Soroka, et al. - 2006 |

8 | Fuzzy clustering with artificial bee colony algorithm. - Karaboga, Ozturk - 2010 |

7 | Computerized analysis of breast lesions in three dimensions using dynamic magnetic-resonance imaging. - KGA, ML, et al. - 1998 |

6 |
Computerized interpretation of breast MRI: investigation of enhancement-variance dynamics.
- Chen, ML, et al.
- 2004
(Show Context)
Citation Context ...ion of the training and testing data, the internal structure of the classifier and the inherent randomness in the training process. The classification performance can be assessed in terms of the sensitivity, specificity and accuracy of the system. Sensitivity (SN) is the proportion of actual positives which are correctly identified and it is mathematically defined (6) and specificity (SP) is the proportion of negatives which are correctly identified and is mathematically defined in (7): Sensitivity: SN = TP TP + FN (6) Specificity: SP = TN TN + FP (7) Accuracy: ACC = TP + TN TP + FP + FN + TN (8) Error in classification: E = 1−ACC, (9) where TP = true positive, TN = true negative, FP = false positive, and FN = false negative. It is obvious that the main objective of a classifier is to minimize the false positive and negative rates, similarly, to maximize the true negative and positive rates. The sensitivity, specificity, accuracy, and error of www.scienceasia.org ScienceAsia 39 (2013) 301 the classification technique were evaluated through quantitative measures derived through the comparison of each classified result with its corresponding ground truth. It is defined in (6)–(9). Groun... |

6 |
Breast lesion analysis of shape technique: Semiautomated versus manual morphological description.
- GP, Sreenivas, et al.
- 2006
(Show Context)
Citation Context ...e internal structure of the classifier and the inherent randomness in the training process. The classification performance can be assessed in terms of the sensitivity, specificity and accuracy of the system. Sensitivity (SN) is the proportion of actual positives which are correctly identified and it is mathematically defined (6) and specificity (SP) is the proportion of negatives which are correctly identified and is mathematically defined in (7): Sensitivity: SN = TP TP + FN (6) Specificity: SP = TN TN + FP (7) Accuracy: ACC = TP + TN TP + FP + FN + TN (8) Error in classification: E = 1−ACC, (9) where TP = true positive, TN = true negative, FP = false positive, and FN = false negative. It is obvious that the main objective of a classifier is to minimize the false positive and negative rates, similarly, to maximize the true negative and positive rates. The sensitivity, specificity, accuracy, and error of www.scienceasia.org ScienceAsia 39 (2013) 301 the classification technique were evaluated through quantitative measures derived through the comparison of each classified result with its corresponding ground truth. It is defined in (6)–(9). Ground truth is based on the diagnosis of the... |

6 | Neural classification of abnormal tissue in digital mammography using statistical features of the texture. In: - Christoyianni, Dermatas, et al. - 1999 |

6 | Classification of breast MRI lesions using a backpropagation neural network (BNN). In: - Arbach, Stolpen, et al. - 2004 |

5 | Multifeature analysis of Gd-enhanced MR images of breast lesions. - Sinha, FA, et al. - 1997 |

4 |
Further signs in the evaluation of magnetic resonance mammography: A retrospective study.
- DR, Wurdinger, et al.
- 2005
(Show Context)
Citation Context ...hin polynomial bounded computation times. The honey bee algorithm is a search algorithm which is capable of locating good solutions efficiently. The algorithm is inspired by the food foraging behaviour of honey bees and could be regarded as belonging to the category of intelligent optimization tools22. In the artificial bee colony (ABC) algorithm, the colony of artificial bees contains three groups of bees: employed bees, onlookers, and scouts. Each cycle of the search consists of three major steps: (1) placing the employed bees onto the food sources and then calculating their nectar amounts; (2) selecting the food sources by the onlookers after sharing the information of employed bees and determining the www.scienceasia.org 296 ScienceAsia 39 (2013) nectar amount of the food sources; (3) determining the scout bees and placing them onto the randomly determined food sources. In the ABC, a food source position represents a possible solution to the problem to be optimized and the nectar amount of a food source corresponds to the quality (fitness) of the associated solution23–25. The ABC algorithm performs a neighbourhood search combined with random search in a way that is indicative of t... |

3 |
Feature selection and classification of breast cancer on dynamic magnetic resonance imaging using genetic algorithm and artificial neural networks.
- Nirooei, Abdolmaleki, et al.
- 2009
(Show Context)
Citation Context ...colony algorithm Many complex multi-variable optimization problems cannot be solved exactly within polynomial bounded computation times. The honey bee algorithm is a search algorithm which is capable of locating good solutions efficiently. The algorithm is inspired by the food foraging behaviour of honey bees and could be regarded as belonging to the category of intelligent optimization tools22. In the artificial bee colony (ABC) algorithm, the colony of artificial bees contains three groups of bees: employed bees, onlookers, and scouts. Each cycle of the search consists of three major steps: (1) placing the employed bees onto the food sources and then calculating their nectar amounts; (2) selecting the food sources by the onlookers after sharing the information of employed bees and determining the www.scienceasia.org 296 ScienceAsia 39 (2013) nectar amount of the food sources; (3) determining the scout bees and placing them onto the randomly determined food sources. In the ABC, a food source position represents a possible solution to the problem to be optimized and the nectar amount of a food source corresponds to the quality (fitness) of the associated solution23–25. The ABC algorit... |

3 | The emerging role of breast magnetic resonance imaging. - AB, RG - 2003 |

3 | Application of the bees algorithm to the training of radial basis function networks for control chart pattern recognition. In: - DT, Ghanbarzadeh, et al. - 2006 |

3 | Registration and lesion classification of contrast-enhanced magnetic resonance breast images. - Tanner - 2005 |

2 |
Approaches for automated detection and classification of masses in mammograms.
- HD, XJ, et al.
- 2006
(Show Context)
Citation Context ...> 0, 1 + |fi |if f < 0. (3) 3: Cycle = 1. 4: Repeat from step 5 to step 8. 5: Apply the training data set to determine the value of the error function associated with each bee. This phase is done by the following process: (i) Produce new solutions Vij in the neighbourhood of Xij for the employed bees using: Vij = Xij + Φij(Xij −Xkj), (4) where k is a solution in the neighbourhood of i, Φ is a random number in the range [−10, 10], and evaluate them. (ii) Apply the Greedy Selection process between processes and calculate the probability values Pi for the solutions Xi as: Pi = fiti∑SN i=1 fiti . (5) (iii) Produce the new solutions Vi for the onlookers from the solutions Xi, selected depending on Pi, and evaluate their fitness. (iv) Calculate the error value between the target and obtained value. 6: Based on the error value obtained from step 5, create a new population of bees comprising the best bees in the selected neighbourhoods and randomly placed scout bees. This phase is done by the following process: (i) Apply the Greedy Selection process for the onlookers between Xi and Vi and determine the abandoned solution (source), if exists, replace it with a new randomly produced solution Xi... |

2 | Gradient vector flow snake segmentation of breast lesions in dynamic contrast-enhanced MR images. In: - Bahreini, Fatemizadeh, et al. - 2010 |

2 | Interactive lesion segmentation on dynamic contrast enhanced breast MRI using a Markov model. - Wu, Salganicoff, et al. - 2006 |

2 | Development of intelligent system based on artificial swarm bee colony clustering algorithm for efficient mass extraction from breast DCE-MR images. - D, Geetha - 2011 |

2 | Comparative study of different edge enhancement filters in spatial domain for magnetic resonance images. - D, Geetha - 2011 |

2 | Introduction to Neural Networks Using Matlab 6.0, Tata McGraw-Hill, - SN, SN - 2006 |

2 | Neural network approach to the segmentation and classification of dynamic magnetic resonance images of the breast: Comparison with empiric and quantitative kinetic parameters. - BK, Aspelin, et al. - 2004 |

2 | Prediction of malignant breast lesions from MRI features: a comparison of artificial neural network and logistic regression techniques. - CE, Chen, et al. - 2009 |

2 | Feature selection and classification of breast cancer on dynamic magnetic resonance imaging using ANN and SVM. - Keyvanfard, MA, et al. - 2011 |

1 | Malignant lesion segmentation in contrast-enhanced breast MR images based on the marker-controlled watershed. - Cui, Tan, et al. - 2009 |

1 | Development of cad system based on enhanced clustering based segmentation algorithm for detection of masses in breast DCEMRI. - D, Geetha - 2011 |

1 | Breast MRI lesion classification: Improved performance of human readers with a backpropagation neural network computer-aided diagnosis (CAD) system. - LA, AH, et al. - 2007 |

1 | Object extraction from lidar data using an artificial swarm bee colony clustering algorithm. In: Stilla U, Rottensteiner F, Paparoditis N (eds) - Saeedi, Samadzadegan, et al. - 2009 |

1 | A bee colony optimization algorithm to job scheduling simulation. In: Perrone - CS, MYH, et al. - 2006 |

1 | Optimization of a fuzzy logic controller using the Bees Algorithm. - DT, A, et al. - 2009 |

1 | Breast tumor analysis using texture features and wavelet transform with dynamic neural network based training. - HK, VD - 2010 |

1 | Breast tissue classification in mammograms using ICA mixture models. In: - Christoyianni, Koutras, et al. - 2001 |

1 | Computer aided mammography. In: - Zhang, Lu, et al. - 2008 |

1 | Elisseeff A - Guyon - 2003 |

1 | Computer aided diagnosis of digital mammograms. In: - WA, YM - 2007 |

1 | Nonmass lesions in magnetic resonance imaging of the breast: additional T2-weighted images improve diagnostic accuracy. - PAT, Dietzel, et al. - 2011 |