Sharifah Masniah Wan Masra

Application of Particle Swarm Optimization in Histogram Equalization for image enhancement

The basic and most common technique for contrast adjustment in the image is using Histogram Equalization (HE). It is based on equalizing the histogram of the image and subsequently enhancing its contrast, and results in overall contrast improvement. This paper introduces a combination of normal HE technique and Particle Swarm Optimization (PSO) algorithm for enhancing distorted image naturally. The process is as follows. The image is separated into red, green and blue (RGB) channels and PSO algorithm is applied into each channel in order to get its best fitness value. The fitness value that is obtained then will be applied into HEs normalization, after that the processed color image will be merged back to RGB image. Experimental results have shown the effectiveness in improving the contrast of the original images without introducing disturbing artifacts caused by normal HE.

An improved binary particle swarm optimization algorithm for DNA encoding enhancement

The accuracy of DNA computing highly depends on the DNA strands used in solving complex computations. As such, many approaches are proposed to design DNA oligonucleotides that are stable and unique. In this paper, an improved binary particle swarm optimization (IBPSO) algorithm is proposed and implemented. Four objective functions which are H-measure, similarity, hairpin and continuity are employed to define the uniqueness of designed sequences. The DNA words are constrained within a predefined range of GC-content and melting temperature. The performances and the ability of the algorithm to enhance the characteristics of generated DNA code words are analyzed. The results obtained show that this algorithm executes better sequences and did perform better compared to other optimization techniques. Moreover, it converges faster than the previously suggested binary particle swarm optimization algorithm.

Sound navigation aid system for the vision impaired

Visually impaired or blind people are someone who lost their ability to see. They usually depend on guide cane or guide dog to assist them navigates their environment. However, guide cane requires the blind to scan the surrounding manually with their hand while guide dog is expensive and has a short life span. There are many researches on creating better navigation aid for blind people. Instead of using vibration or alarm to notify the distance of the obstacles, this research proposed the application of sound notification for the blind. This system is developed using MATLAB r2010a with the integrated webcam of the laptop and a laser pointer. The laser pointer acts as a reference point so that objects distance can be calculated from the captured image by the webcam. This system manages to calculate distance for any color of object under any level of brightness condition during day and night time. It also possesses human head movement ability to turn left and right, up and down. This system managed to detect the obstacles up to a distance of 432 centimeters, which is sufficient to prevent the blind from bumping into the obstacles. The calculated distance is then converted to sound in order to notify the blind how far the object is away from them.

Satellite Image Segmentation with the Application of PSO Algorithm

Satellite imaging consists of capturing images of the Earth through a series of artificial satellites. These images contain an abundance of information that can be used in several applications such as fishing, agriculture, regional planning, biodiversity conservation and many others. Digital image processing can help overcome the limitations of human vision by extracting key information from these images at a much higher rate through the speed of automation. This paper aims to achieve that by exploiting the potential of the Particle Swarm Optimization (PSO) algorithm in image segmentation. Various satellite images were segmented using PSO algorithm before a trace of the objects that have been isolated in the image was run to evaluate the accuracy of segmentation. Three objective measurements which are Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE), were made on the outputs of the segmentation using PSO algorithm and the traditional segmentation technique which is Otsu’s method for comparison. The proposed method which applies the PSO algorithm proved to be superior in producing images of higher quality and accuracy as compared to the traditional segmentation technique.

Ensembles of DNA letters for the design of unique DNA library using a modified version of multi-criteria VEDEPSO optimizer

The DNA code words designing is a multi-criteria combinatorial optimization task. The designed words should be as unique as possible, thermodynamically stable, non-self hybridized, non-cross hybridized with others and have good chemical properties. In this paper, the DNA words designing approach implied concurrent minimizations of four objective functions, H-measure, similarity, hairpin and continuity. The designations is subjected to a predefine range of melting temperature and GC-content. A novel multi-population optimizer, M-VEDEPSO, is employed to design sets of DNA strands. The algorithm runs for 10 times and as a result, each population has lower average fitness values compared to the fitness values obtained using the conventional VEDEPSO algorithm. The results obtained from the algorithm are indicated by 12 randomly selected non-dominated particles/individuals. These solutions are obtained via Pareto dominance concepts.

Development of Portable Finger Clubbing Meter

Finger clubbing, also known as drumstick finger, is the medical symptom that is indicated by the development of the sponginess or swelling in the nail beds of nails and toes. The higher grade of clubbing on the patients can be easily identified with the presence of drumstick finger. The existing available measurement device to identify the early stage of clubbing required much time which is impractical for a busy clinic practice. This paper explains the determination of the finger clubbing by using the Digital Index (DI) measurement, which was deployed by implementing and developing the Portable Finger Clubbing Meter hardware and the Microsoft Visual Basic (VB) Graphical User Interface (GUI). Finger circumference values of nail-fold (NF) and distal interphalangeal joint (DIP) of twenty participants were measured using the developed hardware. Data analysis was then performed using the GUI for DI computation, and the presence of finger clubbing could be determined.

DNA words based on an enhanced algorithm of multi-objective particle swarm optimization in a continuous search space

In this paper, particle swarm optimization algorithm in a continuous search space is implemented to generate a set of DNA words. A single swarm with 20 particles is used to find the best solution (gbest). Here, each particle has a number of sub-particles which is referred as the number of sequences to be designed. Overall, the particle which has the optimum fitness value is considered as the best solution. Fitness value of a particle is computed from the average value of all the applied objective functions. The solution obtained from this algorithm is found to be better as compared with other approaches. Furthermore, it has a fast convergence towards the optimum fitness value.

The effect of input limiting on linear and nonlinear filters for the removal of impulsive noise

This paper investigates the effects of input limiting on the performance of linear and nonlinear filters when the input signal is contaminated by impulsive noise modeled as an alpha-stable (α-stable) random process. The nonlinear filters chosen are from the median family filters: median and recursive median filter, and Wiener filter from the linear filter group. The hypothesis being tested in this paper is that front-end limiting will increased the performance of median filter when the noise was extremely impulsive. Each image has been analyzed for four impulsive noise parameter values, α = 2, 1.5, 1, and 0.5 which cover a range of noise behaviors from Gaussian, through Cauchy, to impulsive. One method of handling very large impulses is to incorporate in the receiver over-voltage protection in the form of an analog limiter which clips the received signal before the filtering process. The simulation results obtained suggest that the front-end limiting enhances the performance of median filters in dealing with images, where it eliminates the highly impulsive noise while preserving useful detail in the image.