Showing papers by "Pradipta Maji published in 2008"

Maximum Class Separability for Rough-Fuzzy C-Means Based Brain MR Image Segmentation

Abstract: Image segmentation is an indispensable process in the visualization of human tissues, particularly during clinical analysis of magnetic resonance (MR) images. In this paper, the rough-fuzzy c -means (RFCM) algorithm is presented for segmentation of brain MR images. The RFCM algorithm comprises a judicious integration of the rough sets, fuzzy sets, and c -means algorithm. While the concept of lower and upper approximations of rough sets deals with vagueness and incompleteness in class definition of brain MR images, the membership function of fuzzy sets enables efficient handling of overlapping classes. The crisp lower bound and fuzzy boundary of a class, introduced in the RFCM algorithm, enable efficient segmentation of brain MR images. One of the major issues of the RFCM based brain MR image segmentation is how to select initial prototypes of different classes or categories. The concept of discriminant analysis, based on the maximization of class separability, is used to circumvent the initialization and local minima problems of the RFCM. Some quantitative indices are introduced to extract local features of brain MR images for accurate segmentation. The effectiveness of the RFCM algorithm, along with a comparison with other related algorithms, is demonstrated on a set of brain MR images.

Second Order Fuzzy Measure and Weighted Co-Occurrence Matrix for Segmentation of Brain MR Images

Abstract: A robust thresholding technique is proposed in this paper for segmentation of brain MR images. It is based on the fuzzy thresholding techniques. Its aim is to threshold the gray level histogram of brain MR images by splitting the image histogram into multiple crisp subsets. The histogram of the given image is thresholded according to the similarity between gray levels. The similarity is assessed through a second order fuzzy measure such as fuzzy correlation, fuzzy entropy, and index of fuzziness. To calculate the second order fuzzy measure, a weighted co-occurrence matrix is presented, which extracts the local information more accurately. Two quantitative indices are introduced to determine the multiple thresholds of the given histogram. The effectiveness of the proposed algorithm, along with a comparisonwith standard thresholding techniques, is demonstrated on a set of brain MR images.

A Novel Biclustering Algorithm for Discovering Value-Coherent Overlapping σ-Biclusters

Abstract: The biclustering method is a very useful tool for analyzing gene expression data when some genes have multiple functions and experimental conditions are diverse in gene expression measurement. It focuses on finding a subset of genes and a subset of experimental conditions that together exhibit coherent behavior. A large number of biclustering algorithms has been developed for analyzing gene expression data. Most of them find exclusive biclusters, which is inappropriate in the biological context. Since biological processes are not independent of each other, many genes participate in multiple different processes. Hence, nonexclusive biclustering algorithms are required for finding highly overlapping biclusters. In this regard, a novel overlapping biclustering algorithm is presented here to find overlapping biclusters of larger volume with mean squared residue lower than a given threshold. The proposed method consists of two phases. First, a set of highly coherent seeds is generated based on two-way k-medoids algorithm, where mutual information is used as a similarity measure instead of using Euclidean distance. The seeds are then iteratively adjusted (enlarged or degenerated) by adding or removing genes and conditions based on a new quantitative index. In effect, the proposed method provides highly overlapping coherent biclusters with mean squared residue lower than a given threshold. Some quantitative indices are introduced for evaluating the quality of generated biclusters. The quality of biclusters found by the proposed approach is discussed and the results are compared to those reported by existing methods. In general, the proposed approach shows an excellent performance at finding patterns in gene expression data.

On Characterization of Attractor Basins of Fuzzy Multiple Attractor Cellular Automata

Abstract: Two new operators, namely, dependency vector (DV) and derived complement vector (DCV) are introduced in this paper to characterize the attractor basins of the additive fuzzy cellular automata (FCA) based associative memory, termed as fuzzy multiple attractor cellular automata (FMACA). The introduction of DV and DCV makes the complexity of the attractor basin identification algorithm linear in time. The characterization of the FMACA using DV and DCV establishes the fact that the FMACA provides both equal and unequal size of attractor basins. Finally, a set of algorithms is proposed to synthesize the FCA rules, attractors, and predecessors of attractors from the given DV and DCV in linear time complexity.

Fault Diagnosis of Electronic Circuits Using Cellular Automata Based Pattern Classifier

Abstract: This chapter formulates fault diagnosis in electronic circuits as a pattern classification problem The proposed pattern classification scheme employs the computing model of a special class of sparse network referred to as cellular automata (CA) A particular class of CA referred to as multiple attractor CA (MACA) has been projected as a classifier of faulty response-pattern of a circuit The genetic algorithm (GA) is employed to synthesize the desired CA required for diagnosis of a circuit under test (CUT) The CUT is assumed to have a network of large number of circuit components partitioned into a number of sub-circuits referred to as modules Introduction of GA significantly reduces the design overhead of the MACA based classifier that supports:

Rough-fuzzy relational clustering algorithm for biological sequence mining

Abstract: This paper presents a hybrid relational clustering algorithm, termed as rough-fuzzy c-medoids, to cluster biological sequences. It comprises a judicious integration of the principles of rough sets, fuzzy sets, c-medoids algorithm, and amino acid mutation matrix used in biology. The concept of crisp lower bound and fuzzy boundary of a class, introduced in rough-fuzzy c-medoids, enables efficient selection of cluster prototypes. The effectiveness of the algorithm, along with a comparison with other algorithms, is demonstrated on different protein data sets.

Pattern Classification Using NNTree: Design and Application for Biological Dataset

Abstract: Recent advancement and wide use of high-throughput technologies for biological research are producing enormous size of biological datasets distributed worldwide. Pattern classification techniques and machine learning methods provide useful tools for knowledge discovery in this field. The goal of this paper is to present the design and application of a neural network tree (NNTree) based pattern classifier to mine biological datasets. An NNTree is a hybrid-learning model with the overall structure being a decision tree (DT), and each non-terminal node containing a neural network (NN). Demon-strating its success in splice-junction and gene identification problems provides the effectiveness of this approach. Extensive experimental results establish that the proposed classifier produces more accurate classifier than that have previously been obtained for a range of different sequence lengths; thereby indicating a cost-effective alternative in splice-junction and protein coding region identification problem.

Prediction of Protein Functional Sites Using Novel String Kernels

Abstract: In most pattern recognition algorithms, amino acids cannot be used directly as inputs since they are nonnumerical variables. They, therefore, need encoding prior to input. In this regard, a novel string kernel is introduced, which maps a nonnumerical sequence space to a numerical feature space.The proposed string kernel is developed based on the conventional bio-basis function and termed as novel bio-basis function. The novel bio-basis function is designed based on the principle of asymmetricity of biological distance, which is calculated using an amino acid mutation matrix. The concept of zone of influence of bio-basis is introduced in the proposed string kernel to normalize the asymmetric distance. An efficient method to select bio-bases for the novel string kernel is described integrating the concepts of the Fisher ratio and degree of resemblance. The effectiveness of the proposed string kernel and bio-bases selection method, along with a comparison with existing kernel and related selection methods, is demonstrated on different protein data sets.