Mohamed Ben Messaoud

Detection of brain tumor in medical images

This paper introduces an efficient detection of brain tumor from cerebral MRI images. The methodology consists of three steps: enhancement, segmentation and classification. To improve the quality of images and limit the risk of distinct regions fusion in the segmentation phase an enhancement process is applied. We adopt mathematical morphology to increase the contrast in MRI images. Then we apply Wavelet Transform in the segmentation process to decompose MRI images. At last, the k-means algorithm is implemented to extract the suspicious regions or tumors. Some of experimental results on brain images show the feasibility and the performance of the proposed approach.

Automated classification of magnetic resonance brain images using Wavelet Genetic Algorithm and Support Vector Machine

In this paper we propose a new approach for automated diagnosis and classification of Magnetic Resonance (MR) human brain images, using Wavelets Transform (WT) as input to Genetic Algorithm (GA) and Support Vector Machine (SVM). The proposed method segregates MR brain images into normal and abnormal. Our contribution employs genetic algorithm for feature selection witch requires much lighter computational burden. An excellent classification rate of 100% could be achieved using the support vector machine. We observe that our results are significantly better than the results reported in a previous research work employing Wavelet Transform and Support Vector Machine.

Medical Image Classification Using an Optimal Feature Extraction Algorithm and a Supervised Classifier Technique

A new approach for automated diagnosis and classification of Magnetic Resonance MR human brain images is proposed. The proposed method uses Wavelets Transform WT as input module to Genetic Algorithm GA and Support Vector Machine SVM. It segregates MR brain images into normal and abnormal. This contribution employs genetic algorithm for feature selection which requires much lighter computational burden in comparison with Sequential Floating Backward Selection SFBS and Sequential Floating Forward Selection SFFS methods. A percentage reduction rate of 88.63% is achieved. An excellent classification rate of 100% could be achieved using the support vector machine. The observed results are significantly better than the results reported in a previous research work employing Wavelet Transform and Support Vector Machine.

Automated segmentation of brain tumor using optimal texture features and support vector machine classifier

This paper presents a new general automatic method for segmenting brain tumors in magnetic resonance (MR) images. Our approach addresses all types of brain tumors. The proposed method involves, subsequently, image pre-processing, feature extraction via wavelet transform (WT), dimensionality reduction using genetic algorithm (GA) and classification of the extracted features using support vector machine (SVM). For the segmentation of brain tumor these optimal features are employed. The resulting method is aimed at early tumor diagnostics support by distinguishing between the brain tissue, benign tumor and malignant tumor tissue. The segmentation results on different types of brain tissue are evaluated by comparison with manual segmentation as well as with other existing techniques.

The bidimensional empirical mode decomposition with 2D-DWT for gaussian image denoising

This paper presents a new adaptive approach for image denoising with Gaussian noise based on a combination of the Bidimensional Empirical Mode Decomposition (BEMD) and the the discrete wavelet transforms (DWT). The BEMD is an auto-adaptive method for the analysis of nonlinear or non-stationary signals and images. The input image is decomposed into several modes called Intrinsic Mode Functions (IMFs), which show new characteristics of the images. In this paper, we propose to apply the BEMD approach in the image denoising domain by using the first IMF to reduce the Gaussian noise in blurred images. After that, we combine the BEMD with the DWT to improve the BEMD denoising method. Finally, we show the influence of the number of IMFs filtered with the DWT on the visual quality in term of PSNR of the denoised image.

Frequency recognition based on the Inter-Battery Factor Analysis for SSVEP-based BCIs

The Brain-Computer Interface is a system mainly designed to provide people suffering from severe neuromuscular disorder with a new mean of communication and control. Increasing the system accuracy rate is the goal of several studies. In fact, ameliorating this criterion allows to minimize the correction phase and makes the use of the system more natural. This is very important to develop Brain-Computer interface systems for everyday use outside the laboratory. This paper introduces a new Brain-Computer Interface based on the Inter-Battery Factor Analysis method. The results show that the proposed BCI system has a higher accuracy than systems based on Canonical Correlation Analysis or Multivariate Synchronization Index. The accuracy rate has reached 93.2% for the five participants using only the two electrodes O1 and O2 with data acquired over a period of 2.25s.

Anti-noise Capability Improvement of Minimum Energy Combination Method for SSVEP Detection

Minimum energy combination (MEC) is a widely used method for frequency recognition in steady state visual evoked potential based BCI systems. Although it can reach acceptable performances, this method remains sensitive to noise. This paper introduces a new technique for the improvement of the MEC method allowing ameliorating its Anti-noise capability. The Empirical mode decomposition (EMD) and the moving average filter were used to separate noise from relevant signals. The results show that the proposed BCI system has a higher accuracy than systems based on Canonical Correlation Analysis (CCA) or Multivariate Synchronization Index (MSI). In fact, the system achieves an average accuracy of about 99% using real data measured from five subjects by means of the EPOC EMOTIVE headset with three visual stimuli. Also by using four commands, the system accuracy reaches 91.78% with an information-transfer rate of about 27.18 bits/min.

Image Denoising In Gaussian and Impulsive Noise Based On Block Bidimensional Empirical Mode Decomposition

In this paper we develop an adaptive algorithm for decomposition and filtering of grayscales images. This method is highly adaptive decomposition image called Bidimensional Empirical Mode Decomposition (BEMD) based in blocks. This proposed approach decomposes image into a basis functions named Intrinsic Mode Function (IMF) and residue. This method offers a good result in visual quality but it consumes an important execution time. To overcome this problem we propose a new approach using Block based BEMD method where the input image is subdivided into blocks. Then the conventional BEMD is applied on each of the four blocks separately. This proposed extended method gives a solution in reduction of execution time. This approach shows the good results in the field of image filtering. Denoised image is obtained by summing the residue and the filtered first IMFs (the detail) using a wavelet technique. Experimental results positively show that this proposed methodology removes Gaussian and Impulsive noises from the images. General Terms Image filtering, image decomposition.

Image decomposition based on modified bidimensional empirical mode decomposition

In this paper we develop an adaptive algorithm for decomposition of greyscales images. This method is highly adaptive decomposition image called Bidimentional Empirical Mode Decomposition (BEMD). It is based on the characterization of the image through its decomposition in Intrinsic Mode Function (IMF) where it can be decomposed into basis functions called IMF and a residue. This method offered a good result in visual quality, unfortunately this method consume an important execution time. To overcome this problem we proposed a new approach using Block based BEMD method where the input image is subdivided into blocks. Then the BEMD is applied on each of the four blocks separately. This method offered a good solution to reduce the execution time.

Bispectral analysis-based approach for steady-state visual evoked potentials detection

Brain-Computer Interface (BCI) systems are widely based on steady-state visual evoked potentials (SSVEP) detection using electroencephalography (EEG) signals. SSVEP-based BCIs are becoming attractive due to their higher signal-to-noise ratio (SNR) as well as faster information transfer rate (ITR). However, their performances are largely affected by the interference coming from the spontaneous EEG activities which intrinsically restrict their efficiency in distinguishing between SSVEPs and background EEG activities. In this paper, we introduce a new approach for the detection of SSVEP based on bispectral analysis to palliate the frequency-dependent bias. A COMB filter associated with a wavelet denoising filter is firstly used to minimize the noise while improving the SNR of phase signals. Next, the complementary orthogonal projections and the principle component analysis (PCA) are used to decompose the components related to SSVEPs and components related to brain activities. Finally, the bispectrum, a powerful tool for the analysis and the characterization of nonlinear properties of stochastic signals, is used to extract the features of the EEG signal benefiting from the information about the phase coupling of the signal components. The results of experiments, using two databases on five (or ten) subjects, show that the proposed approach significantly outperformed the standard CCA approach in distinguishing the target frequency and in average information transfer rate.