Amir Tahmasbi

Classification of benign and malignant masses based on Zernike moments

In mammography diagnosis systems, high False Negative Rate (FNR) has always been a significant problem since a false negative answer may lead to a patients death. This paper is directed towards the development of a novel Computer-aided Diagnosis (CADx) system for the diagnosis of breast masses. It aims at intensifying the performance of CADx algorithms as well as reducing the FNR by utilizing Zernike moments as descriptors of shape and margin characteristics. The input Regions of Interest (ROIs) are segmented manually and further subjected to a number of preprocessing stages. The outcomes of preprocessing stage are two processed images containing co-scaled translated masses. Besides, one of these images represents the shape characteristics of the mass, while the other describes the margin characteristics. Two groups of Zernike moments have been extracted from the preprocessed images and applied to the feature selection stage. Each group includes 32 moments with different orders and iterations. Considering the performance of the overall CADx system, the most effective moments have been chosen and applied to a Multi-layer Perceptron (MLP) classifier, employing both generic Back Propagation (BP) and Opposition-based Learning (OBL) algorithms. The Receiver Operational Characteristics (ROC) curve and the performance of resulting CADx systems are analyzed for each group of features. The designed systems yield Az=0.976, representing fair sensitivity, and Az=0.975 demonstrating fair specificity. The best achieved FNR and FPR are 0.0% and 5.5%, respectively.

Fast opposite weight learning rules with application in breast cancer diagnosis

Classification of breast abnormalities such as masses is a challenging task for radiologists. Computer-aided Diagnosis (CADx) technology may enhance the performance of radiologists by assisting them in classifying patterns into benign and malignant categories. Although Neural Networks (NN) such as Multilayer Perceptron (MLP) have drawbacks, namely long training times, a considerable number of CADx systems employ NN-based classifiers. The reason being that they provide high accuracy when they are appropriately trained. In this paper, we introduce three novel learning rules called Opposite Weight Back Propagation per Pattern (OWBPP), Opposite Weight Back Propagation per Epoch (OWBPE), and Opposite Weight Back Propagation per Pattern in Initialization (OWBPI) to accelerate the training procedure of an MLP classifier. We then develop CADx systems for the diagnosis of breast masses employing the traditional Back Propagation (BP), OWBPP, OWBPE and OWBPI algorithms on MLP classifiers. We quantitatively analyze the accuracy and convergence rate of each system. The results suggest that the convergence rate of the proposed OWBPE algorithm is more than 4 times faster than that of the traditional BP. Moreover, the CADx systems which use OWBPE classifier on average yield an area under Receiver Operating Characteristic (ROC), i.e. Az, of 0.928, a False Negative Rate (FNR) of 9.9% and a False Positive Rate (FPR) of 11.94%.

A novel opposition-based classifier for mass diagnosis in mammography images

In this paper, a novel opposition-based classifier has been developed which classifies breast masses into benign and malignant categories. An MLP network with a novel learning rule, called Opposite Weighted Back Propagation (OWBP), has been utilized as the classifier. The objective is increasing the convergence rate of MLP learning rules as well as improving the mass diagnostic performance. The input ROI, which is a suspected part of mammogram, is segmented manually by expert radiologists and subjected to some preprocessing stages such as histogram equalization, translation and scaling. Then, a group of features which are appropriate descriptors of mass shape, margin and density have been extracted from the preprocessed ROIs. The proposed features include circularity, Zernike moments, contrast, average gray level, NRL derivatives and SP. The proposed classifier has been trained with both traditional BP and OWBP learning rules and the performance have been evaluated. The system which utilizes OWPB learning rule yields a significantly faster training time than BP algorithm while the Az of the resulting CADx system is 0.944.

Mass diagnosis in mammography images using novel FTRD features

In this paper, a novel group of features have been introduced for diagnosing the masses in mammography images. The goal is increasing the performance of CADx algorithms as well as decreasing computational complexity. The proposed features are proper descriptors of mass margin which are called Fourier Transform of Radial Distance (FTRD). The input ROI has been segmented manually by expert radiologists and subjected to some preprocessing stages. In order to extract the proposed features, the Radial Distance (RD) vectors of masses have been extracted. In addition, the zero padding method has been utilized to equalize the length of the RD vectors. Then, the resulting vectors are transformed to the frequency domain. It is shown that the magnitude response of FTRD vectors can be appropriate descriptors of the mass margin. Furthermore, in order to make a trade-off between the computational complexity and performance of the overall system, several groups of FTRD features with different lengths have been chosen and applied to an MLP classifier. Finally, the ROC curves have been plotted for each group of features and the performances have been evaluated. The most effective system yields an Az which is equal to 0.98. Moreover, the best achieved FPR is 5.56%.

An effective breast mass diagnosis system using Zernike moments

In this paper, a novel CADx system has been proposed for the diagnosis of masses in mammography images. The objective is intensifying the performance of CADx algorithms as well as reducing the false positive rate by utilizing Zernike moments as descriptors of shape and margin characteristics. The input ROI is segmented manually by expert radiologists. Then, it is subjected to some preprocessing stages such as histogram equalization, translation, and NRL scaling. The outcome of preprocessing stage is two processed images containing co-scaled translated masses. Besides, one of these images represents the shape characteristics of the mass, while the other describes the margin characteristics. Two groups of Zernike moments have been extracted from the preprocessed images and proceeded to the feature selection stage. Each group includes 32 moments with different orders and iterations. Considering the performance of the overall CADx system, the most effective 32 moments have been chosen and applied to a multi-layer Perceptron classifier. The ROC plot and the performance of overall CADx system are analyzed for each group of features. The designed systems yield Az = 0.976 and 0.975 which represent fair sensitivity and fair specificity, respectively. The best achieved FPR is 5.5%.

Classification of Breast Masses based on Cognitive Resonance

In this pape r, a novel approach has been proposed for mass diagnosis in mammography images. The objective is developing a trustable mammography Computer Aided-Diagnosis (CADx) system utilizing a new cognitive classifier. The input Region of Interest (ROI) is subjected to some preprocessing stages; then, a group of features describing the shape, margin and density characteristics of masses have been extracted. The proposed features are consistent with the evaluations that an expert radiologist takes into account in diagnosis process. The most effective features are selected in the feature selection stage and mapped from the set of real numbers to a set of linguistic terms. The proposed classifier primes a knowledge-base which is developed according to a mammography expert; its rules have been written using a special kind of linguistics and grammar formalism. The semantic comparison of features of the image to the expectations of the knowledge base, called cognitive resonance, leads to the final assessment. Since the output of this system comes with reason, the system is trustable. The best achieved Accuracy and False Positive Rate (FPR) are 87.93% and 10.52%, respectively. More numerical results are reported in the paper.

CWLA: A novel cognitive classifier for breast mass diagnosis

A novel cognitive classifier has been introduced to develop a trustable mammography Computer Aided Diagnosis (CADx) system which is called Cognitive Weighted Linear Aggregation (CWLA). A group of in-depth analyzed features are extracted from the preprocessed Regions of Interest (ROIs) and mapped from set of real numbers to a set of linguistic terms. The proposed classifier primes a knowledge base which is developed according to a mammography expert. The semantic comparison of the extracted features with the expectations of the knowledge base, which is called cognitive resonance, leads to a primary clustering. Finally, the linguistic terms are remapped onto the set of real numbers and the final assessment comes out from the weighted linear aggregation of clustered categories. Since the output of the system comes with reason, the system is reliable. The achieved area under Receiver Operational Characteristics (ROC) curve (Az) and False Positive Rate (FPR) are 0.858 and 5.26%, respectively.

New Optimized IIR Low-Pass Differentiators

In this paper a novel approach is proposed for approximating Parks-McClellan low-pass differentiators using optimized low-order IIR filters. Indeed, a suitable IIR filter is designed for approximating Parks- McClellan Low pass differentiator using modified Al-Alaoui’s method, and then denominator polynomial coefficients of resulting transfer function optimized by Genetic algorithm. A suitable fitness function is defined to optimize both magnitude and phase responses; moreover, appropriate weighting coefficients and GA parameters are reported for several cut-off frequencies. It is shown that the order-4 proposed low-pass differentiators yield a frequency response which is almost equal to order-30Parks-McClellan low-pass differentiators. Furthermore, they yield steep roll-off properties, small magnitude error and almost linear phase in the pass-band; the percentage error of magnitude response is less than 0.5%.

Using Symlet Decomposition Method, Fuzzy Integral and Fisherface Algorithm for Face Recognition

In this paper, we have proposed an approach for face recognition by composing Symlet decomposition, Fisherface algorithm, and Sugeno and Choquet Fuzzy Integral. This approach consists of four main sections: the first section uses Symlet, one of the Wavelet families, to transform an image into four sub-images which are called approximate, horizontal, vertical and diagonal partial images respectively. The aim of this work is to extract intrinsic facial features. The second section of this approach uses Fisherface method which is composed of PCA and LDA. The reason for using this was that it is not sensitive to intensive light variations and facial expression and gesture. The third and forth section of this approach, are related to the aggregation of the individual classifiers by means of the fuzzy integral. Both Sugeno and Choquet fuzzy integral are considered as methods for classifier aggregation. In this paper, Olivetti Research Labs face database is used for acquiring experimental results. The approach presented in this paper, will lead to better classification performance compared to other classification methods.

Using subclass discriminant analysis, fuzzy integral and symlet decomposition for face recognition

In this paper, an approach has been proposed for face recognition by composing symlet decomposition, Subclass Discriminant Analysis (SDA), and Sugeno and Choquet Fuzzy Integral. This approach consists of four main sections: the first section uses Symlet, one of the Wavelet families, to transform an image into four sub-images which are called approximate, horizontal, vertical and diagonal partial images respectively. The aim of this work is to extract intrinsic facial features. The second section is composed of PCA and SDA. The reason for using this method was the fact that it is not sensitive to intensive light variations. The third and forth section of this paper are related to the aggregation of the individual classifiers by means of the fuzzy integral. Both Sugeno and Choquet fuzzy integral are considered as methods for classifier aggregation. In this paper, Olivetti Research Labs face database is used for acquiring experimental results. The combination of this method with sym decomposition has yielded a recognition rate which is equal to 97.5 %. The approach presented in this paper, will lead to better classification performance compared to other classification methods.