Chikkannan Eswaran

Approximate Entropy-Based Epileptic EEG Detection Using Artificial Neural Networks

The electroencephalogram (EEG) signal plays an important role in the diagnosis of epilepsy. The EEG recordings of the ambulatory recording systems generate very lengthy data and the detection of the epileptic activity requires a time-consuming analysis of the entire length of the EEG data by an expert. The traditional methods of analysis being tedious, many automated diagnostic systems for epilepsy have emerged in recent years. This paper proposes a neural-network-based automated epileptic EEG detection system that uses approximate entropy (ApEn) as the input feature. ApEn is a statistical parameter that measures the predictability of the current amplitude values of a physiological signal based on its previous amplitude values. It is known that the value of the ApEn drops sharply during an epileptic seizure and this fact is used in the proposed system. Two different types of neural networks, namely, Elman and probabilistic neural networks, are considered in this paper. ApEn is used for the first time in the proposed system for the detection of epilepsy using neural networks. It is shown that the overall accuracy values as high as 100% can be achieved by using the proposed system

An automated decision-support system for non-proliferative diabetic retinopathy disease based on MAs and HAs detection

Diabetic retinopathy (DR) has become a serious threat in our society, which causes 45% of the legal blindness in diabetes patients. Early detection as well as the periodic screening of DR helps in reducing the progress of this disease and in preventing the subsequent loss of visual capability. This paper provides an automated diagnosis system for DR integrated with a user-friendly interface. The grading of the severity level of DR is based on detecting and analyzing the early clinical signs associated with the disease, such as microaneurysms (MAs) and hemorrhages (HAs). The system extracts some retinal features, such as optic disc, fovea, and retinal tissue for easier segmentation of dark spot lesions in the fundus images. That is followed by the classification of the correctly segmented spots into MAs and HAs. Based on the number and location of MAs and HAs, the system quantifies the severity level of DR. A database of 98 color images is used in order to evaluate the performance of the developed system. From the experimental results, it is found that the proposed system achieves 84.31% and 87.53% values in terms of sensitivity for the detection of MAs and HAs respectively. In terms of specificity, the system achieves 93.63% and 95.08% values for the detection of MAs and HAs respectively. Also, the proposed system achieves 68.98% and 74.91% values in terms of kappa coefficient for the detection of MAs and HAs respectively. Moreover, the system yields sensitivity and specificity values of 89.47% and 95.65% for the classification of DR versus normal.

Using genetic algorithm to select the presentation order of training patterns that improves simplified fuzzy ARTMAP classification performance

The presentation order of training patterns to a simplified fuzzy ARTMAP (SFAM) neural network affects the classification performance. The common method to solve this problem is to use several simulations with training patterns presented in random order, where voting strategy is used to compute the final performance. Recently, an ordering method based on min-max clustering was introduced to select the presentation order of training patterns based on a single simulation. In this paper, another single simulation method based on genetic algorithm is proposed to obtain the presentation order of training patterns for improving the performance of SFAM. The proposed method is applied to a 40-class individual classification problem using visual evoked potential signals and three other datasets from UCI repository. The proposed method has the advantages of improved classification performance, smaller network size and lower training time compared to the random ordering and min-max methods. When compared to the random ordering method, the new ordering scheme has the additional advantage of requiring only a single simulation. As the proposed method is general, it can also be applied to a fuzzy ARTMAP neural network when it is used as a classifier.

Performance Evaluation of Neural Network and Linear Predictors for Near-Lossless Compression of EEG Signals

This paper presents a comparison of the performances of neural network and linear predictors for near-lossless compression of EEG signals. Three neural network predictors, namely, single-layer perceptron (SLP), multilayer perceptron (MLP), and Elman network (EN), and two linear predictors, namely, autoregressive model (AR) and finite-impulse response filter (FIR) are used. For all the predictors, uniform quantization is applied on the residue signals obtained as the difference between the original and the predicted values. The maximum allowable reconstruction error delta is varied to determine the theoretical bound delta0 for near-lossless compression and the corresponding bit rate rp. It is shown that among all the predictors, the SLP yields the best results in achieving the lowest values for delta0 and rp. The corresponding values of the fidelity parameters, namely, percent of root-mean-square difference, peak SNR and cross correlation are also determined. A compression efficiency of 82.8% is achieved using the SLP with a near-lossless bound delta0=3, with the diagnostic quality of the reconstructed EEG signal preserved. Thus, the proposed near-lossless scheme facilitates transmission of real time as well as offline EEG signals over network to remote interpretation center economically with less bandwidth utilization compared to other known lossless and near-lossless schemes.

An Automated Blood Vessel Segmentation Algorithm Using Histogram Equalization and Automatic Threshold Selection

This paper focuses on the detection of retinal blood vessels which play a vital role in reducing the proliferative diabetic retinopathy and for preventing the loss of visual capability. The proposed algorithm which takes advantage of the powerful preprocessing techniques such as the contrast enhancement and thresholding offers an automated segmentation procedure for retinal blood vessels. To evaluate the performance of the new algorithm, experiments are conducted on 40 images collected from DRIVE database. The results show that the proposed algorithm performs better than the other known algorithms in terms of accuracy. Furthermore, the proposed algorithm being simple and easy to implement, is best suited for fast processing applications.

An Adaptive And Parameterized Job Grouping Algorithm For Scheduling Grid Jobs

An unorganized deployment of grid applications with a large amount of fine-grain jobs would let the communication overhead dominate the overall processing time, resulting in a low computation-communication ratio. Grids dynamic nature complicates the planning of the job scheduling activity for minimizing the application processing time. This paper presents a grid job scheduling algorithm, based on a parameterized job grouping strategy, which is adaptive to the runtime grid environment. Jobs are grouped based on the job processing requirements, resource policies, network conditions and users QoS requirements. Simulations using the GridSim toolkit reveal that the algorithm reduces the overall application processing time significantly.

Extraction of the Contours of Optic Disc and Exudates Based on Marker-Controlled Watershed Segmentation

The optic disc (OD) and exudates form the main features of fundus images for diagnosing eye disease such as diabetic retinopathy and glaucoma. In this paper, an algorithm for the extraction of OD and exudates from fundus images based on marker controlled watershed segmentation is presented. The proposed algorithm makes use of average filtering and contrast adjustment as preprocessing steps before the watershed transformation is applied. The performance of the proposed algorithm is evaluated using the test images of STARE and DRIVE databases. The results are compared with those reported earlier in the literature. It is shown that the proposed method can yield an average sensitivity value of 94%, which is higher than the value reported earlier.

An enhanced hybrid method for time series prediction using linear and neural network models

The need for improving the accuracy of time series prediction has motivated researchers to develop more efficient prediction models. The accuracy rates resulting from linear models such as linear regression (LR), exponential smoothing (ES) and autoregressive integrated moving average (ARIMA) are not high as they are poor in handling the nonlinear time series data. Neural network models are considered to be better in handling such nonlinear time series data. In the real-world problems, the time series data consist of complex linear and nonlinear patterns and it may be difficult to obtain high prediction accuracy rates using only linear or neural network models. Hybrid models which combine both linear and neural network models can be used to obtain high prediction accuracy rates. In this paper, we propose an enhanced hybrid model which indicates for a given input data which choice is better between the two options, namely, a linear-nonlinear combination or a nonlinear-linear combination. The appropriate combination is selected based on a linearity test of data. From the experimental results, it is found that the proposed hybrid model comprising linear-nonlinear combination performs better than other models for the data that have a linear relationship. On the contrary, the hybrid model comprising nonlinear-linear combination performs better than other models for the data that have a nonlinear relationship.

Exemplar extraction using spatio-temporal hierarchical agglomerative clustering for face recognition in video

Many recent works have attempted to improve object recognition by exploiting temporal dynamics, an intrinsic property of video sequences. In this paper, a new spatio-temporal hierarchical agglomerative clustering (STHAC) method is proposed for automatic extraction of face exemplars for face recognition in video sequences. Two variants of STHAC are presented — a global variety that unifies spatial and temporal distances between points, and a local variety that introduces perturbation of distances based on a local spatio-temporal neighborhood criterion. Faces that are nearest to the cluster means are chosen as exemplars for the testing stage, where subjects in the test video sequences are recognized using a probabilistic-based classifier. Extensive evaluation on a face video database demonstrates the effectiveness of our proposed method, and the significance of incorporating temporal information for exemplar extraction.

Using Autoencoders for Mammogram Compression

This paper presents the results obtained for medical image compression using autoencoder neural networks. Since mammograms (medical images) are usually of big sizes, training of autoencoders becomes extremely tedious and difficult if the whole image is used for training. We show in this paper that the autoencoders can be trained successfully by using image patches instead of the whole image. The compression performances of different types of autoencoders are compared based on two parameters, namely mean square error and structural similarity index. It is found from the experimental results that the autoencoder which does not use Restricted Boltzmann Machine pre-training yields better results than those which use this pre-training method.