N. Md Tahir

Classification of intelligence quotient via brainwave sub-band power ratio features and artificial neural network

This paper elaborates on the novel intelligence assessment method using the brainwave sub-band power ratio features. The study focuses only on the left hemisphere brainwave in its relaxed state. Distinct intelligence quotient groups have been established earlier from the score of the Raven Progressive Matrices. Sub-band power ratios are calculated from energy spectral density of theta, alpha and beta frequency bands. Synthetic data have been generated to increase dataset from 50 to 120. The features are used as input to the artificial neural network. Subsequently, the brain behaviour model has been developed using an artificial neural network that is trained with optimized learning rate, momentum constant and hidden nodes. Findings indicate that the distinct intelligence quotient groups can be classified from the brainwave sub-band power ratios with 100% training and 88.89% testing accuracies.

Brainwave sub-band power ratio characteristics in intelligence assessment

This paper discusses on the brainwave sub-band characteristics for different intelligence groups based on electroencephalogram (EEG) power ratio technique. The EEG datasets have been collected from 50 healthy subjects for two sessions; at relaxed, closed eye (CE) state as reference and at cognitively-stimulated state. In the stimulated state, subjects need to answer the intelligence quotient (IQ) test based on Ravens Standard Progressive Matrices (RPM). Sub-band power ratio from the two sessions were calculated and further analyzed to observe the pattern among different IQ groups. The results show that by implementing power ratio technique, the pattern of IQ groups, especially in the relaxed state can be clearly observed. It can be concluded that the value for alpha ratio is higher for high IQ group compared to low IQ group. In contrast to beta and theta ratio where high IQ groups have lower value compared to the low IQ group. This indicates that the ESD ratios can discriminate the characteristic of brainwaves for intelligence assessment.

EEG sub-band spectral centroid frequencies extraction based on Hamming and equiripple filters: A comparative study

The paper discusses on the effects of Hamming and equiripple filters in the extraction of EEG sub-band spectral centroid frequencies. A total of 40 healthy male subjects have participated in the study. EEG signal sub-bands were filtered using Hamming and equiripple filters with similar frequency response characteristics. It has been observed that the mean difference and variance is very small for all major EEG sub-bands. Implementation of Hamming filter however, induced 91% higher computational requirements as compared to that of the equiripple filters. Hence findings reveal that with similar frequency response characteristics, equiripple filters perform as the more efficient alternative for computation of spectral centroid frequencies.

New image enhancement technique for WMH segmentation of MRI FLAIR image

This paper proposes a new image enhancement technique known as Average Intensity Replacement based on Adaptive Histogram Equalization (AIR-AHE) for FLAIR image based on intensities and contrast mapping techniques. The proposed algorithm consists of partial contrast stretching, contrast limiting enhancement, window sliding neighborhood operation and new pixel centroid replacement. The fluid attenuated inversion recovery (FLAIR) sequences of MRI images which are used for segmentation have low contrast. Therefore, contrast stretching is used to improve the quality of the image. After improving the quality of image, the regions of high intensity are determined to represent potential WMH areas. The result shows that the image has a moderate enhancement on the WMH region which is significant to the image contrast enhancement. With complete brightness preservation, the proposed method gives a relatively natural brightness improvement on the WMH of the periventricular region.

Object detection using depth information from Kinect sensor

This paper investigates the detection of object using depth information that obtained from Kinect sensor. Type and distance of object have been investigated in order to find the best parameter in detecting objects. Two types of object are experimented; shiny and curved shape of object and dull and flat shape of object. Various distances of objects from the Kinect have been tested. From the results obtained, incorrect detection of object has been occurred when using object with shiny and curve shape. It also shows that the optimum distance of object from the Kinect is in the range of 4 feet to 5 feet.

Assessing intensity of white matter hyperintensity and normal appearing white matter in healthy adults

There have been interest on white matter hyperintensity (WMH) and normal white matter (WM) changes reported but have not yet been fully characterized. Different image sequences of magnetic resonance imaging (MRI) scans may shows different gray scale intensity. However, it is difficult to differentiate the intensity of normal WM and WMH as their intensities are visually not much different. In this study, normal WM and WMH changes were investigated based on their intensity to determine the correlation of WMH types and severity in brain of healthy subjects. The assessment was performed by using fully automatic WMH detection and computing algorithms. The main brain regions were segregated into gray matter (GM), normal WM, cerebrospinal fluid (CSF) and non-brain tissue. From the results, it shows that there was significant difference seen between normal appearing WM and hyperintense WM in terms of their intensity levels. The study shows that the development of WMH is prevalent to the occasion of normal WM changes. This is shows that WMH intensity reflects the level of WMH classes and severity; however, further investigations are needed to improve their efficiency.

EEG spectral centroid amplitude and band power features: A correlation analysis

The paper elaborates on the correlation analysis between the proposed EEG sub-band spectral centroid amplitude with the established band power features. The study involves recording of resting EEG from 40 healthy university students. Initially, the EEG is pre-processed for noise removal and filtered into delta, theta, alpha and beta waves using band-pass filters. Next, the sub-band power spectral densities are estimated via Welch method. Subsequently, spectral centroid amplitude and band power features are then computed from the power spectral density of the respective sub-bands. Correlation analysis between the two features is performed for each scalp location to ascertain its relationship. Findings have revealed that the EEG sub-band spectral centroid amplitude is highly correlated with the band power features for all scalp locations.

A New Technique for K-Means Cluster Centers Initialization of WMH Segmentation

K-means algorithm is the most common clustering algorithm being used in medical image processing application. However, the performance of k-means clustering algorithms which converges to numerous local minima would rely on the best initial cluster centers. Generally initial cluster centers are selected randomly and the results are varying on different runs of the algorithm on the same dataset. In this paper, a new method for selecting the best initial centers of k-means clustering is proposed for grouping brain tissues of MRI images. The selection of initial cluster centers namely as Gray Scale Region Intensities (GSRI), is made based on the average intensity value of grayscale region of the images. The proposed method is compared to other method namely as Gray Scale Division Equality (GSDE) which the initial centers were computed by dividing the gray scale 255 and number of clusters. The results show that GSRI outperformed GSDE method in terms of refined segmented regions and converge to local minima with higher iteration number. As a conclusion, it is observed that the newly proposed method has good performance to obtain the initial cluster centers for the k-means algorithm.