Ahmed T. Sahlol

Computer-Aided Acute Lymphoblastic Leukemia Diagnosis System Based on Image Analysis

Leukemia is a kind of cancer that basically begins in the bone marrow. It is caused by excessive production of leukocytes that replace normal blood cells. This chapter presents Computer-Aided Acute Lymphoblastic Leukemia (ALL) diagnosis system based on image analysis. It presented to identify the cells ALL by segmenting each cell in the microscopic images, and then classify each segmented cell to be normal or affected. A well-known dataset was used in this chapter (ALL-IDB2). The dataset contains 260 cell images: 130 normal and 130 affected by ALL. The proposed system starts by segmenting the white blood cells. This process includes sub-processes such as conversion from RGB to CMYK color model, histogram equalization, thresholding by Zack technique, and background removal operation. Then some features were extracted from each cell, each of them represents aspects of a cell. The extracted features include color, texture, and shape features. Then each feature set was exposed to three data normalization techniques z-score, min-max, and grey-scaling to narrow down the gap between the features values. Finally, different classifiers were used to validate the proposed system. The proposed diagnosing system achieved acceptable accuracies when tested by well-known classifiers; however, K-NN achieved the best classification accuracy.

Multimodal Biometric Personal Identification and Verification

Security systems using one identification tool are not ideal. Multisystem security, which using two or more types of security levels like for example using identification password and card, can increase the security of a system, however it is not an ideal security system. Password maybe hacked or forgotten, and Identification card is something we have and could be stolen. This chapter proposes a cascaded multimodal biometric system using fingerprint and iris recognition based on minutiae extraction for fingerprint identification and encoding the log-Gabor filtering for iris recognition. The experiments compare FAR, FRR, and accuracy evaluation metrics for a unimodal biometric system based on either fingerprint or iris and the cascaded multimodal biometric system that sequentially utilizes the fingerprint and iris traits. The proposed system has FAR = 0, FRR = 0.057, and accuracy 99.86%. The results show the superior performance of the proposed multimodal system compared to the unimodal system.

Investigating of Preprocessing Techniques and Novel Features in Recognition of Handwritten Arabic Characters

There are many difficulties facing a handwritten Arabic recognition system such as unlimited variation in character shapes. This paper describes a new method for handwritten Arabic character recognition. We propose a novel efficient approach for the recognition of off-line Arabic handwritten characters. The approach is based on novel preprocessing operations, structural statistical and topological features from the main body of the character and also from the secondary components. Evaluation of the importance and accuracy of the selected features was made. Our method based on the selected features and the system was built, trained and tested by CENPRMI dataset. We used SVM (RBF) and KNN for classification to find the recognition accuracy. The proposed algorithm obtained promising results in terms of accuracy; with recognition rates of 89.2% for SVM. Compared with other related works and also our recently published work we find that our result is the highest among them.

Training feedforward neural networks using Sine-Cosine algorithm to improve the prediction of liver enzymes on fish farmed on nano-selenite

Analytical prediction of oxidative stress biomarkers in ecosystem provides an expressive result for many stressors. These oxidative stress biomarkers including superoxide dismutase, glutathione peroxidase and catalase activity in fish liver tissue were analyzed within feeding different levels of selenium nanoparticles. Se-nanoparticles represent a salient defense mechanism in oxidative stress within certain limits; however, stress can be engendered from toxic levels of these nanoparticles. For instance, prediction of the level of pollution and/or stressors was elucidated to be improved with different levels of selenium nanoparticles using the bio-inspired Sine-Cosine algorithm (SCA). In this paper, we improved the prediction accuracy of liver enzymes of fish fed by nano-selenite by developing a neural network model based on SCA, that can train and update the weights and the biases of the network until reaching the optimum value. The performance of the proposed model is better and achieved more efficient than other models.

Prediction of Antioxidant Status in Fish Farmed on Selenium Nanoparticles using Neural Network Regression Algorithm

Oxidative stress is the most common stress form which is responsible for the increased mortality as well as retardation of productivity in sheries. Selenium plays a vital role in combating oxidative stress. It appears as a potent antioxidant with reduced toxicity in its nanoscale form. In this paper, the effect of the different concentrations of Nano-selenium in the diet on the antioxidant status of common carp was investigated through the estimation of antioxidant enzymes activity and some biochemical blood prole. The adopted regression algorithm for prediction was Back-propagation Neural Network. The model compromised between fast analytical technologies and biological aspect through prediction the healthy status and expected hazards related to oxidative stress. The experiment was performed on four groups of common carp with measured rearing parameters and the same amount of diet at the rates of 0 (control), 0.5, 1 and 2 mg/k gm amount of Nano-selenium concentration in the ration, aiming to build preliminary prediction models to know the antioxidant status activity. The regression performance was tested by several mathematical validations including MSE (Mean squared error), RMSE (Root mean squared error), MSRE (mean squared relative error), MARE (mean absolute relative error), RMSRE (root mean squared relative error), MAE (Mean absolute error), MAPE (Mean absolute percentage error), MSPE (mean squared percentage error), RMSPE (root mean squared percentage error) which showed promising results of the regression model.

Bio-inspired BAT optimization algorithm for handwritten Arabic characters recognition

There are many difficulties facing a handwritten Arabic recognition system such as unlimited variation in human handwriting, similarities of distinct character shapes, interconnections of neighboring characters and their position in the word. This paper presents a handwritten Arabic character recognition system based on BA algorithm. BA algorithm is adopted to reduce the feature set size and to improve the accuracy rate. The proposed system is trained and tested by four well-known classifiers; Bayes Network (BN), artificial neural network (ANN), K-nearest neighbors (KNN), and Random forest (RF) with CENPARMI dataset. The proposed optimization algorithm obtained promising results in terms of classification accuracy as the proposed system is able to recognize 91.59 % of our test set correctly, as well as in terms of computational time reduction. BA algorithm is more efficient in most experiments when comparing with GA and PSO. When compared our results with other related works we find that our result is the highest among other published results.

Particle Swarm Optimization with Random Forests for Handwritten Arabic Recognition System

There are many problems facing the processing of a handwritten Arabic recognition system such as unlimited variation in human handwriting, similarities of character shapes and their position in the word. This paper presents a handwritten Arabic character recognition system based on Particle Swarm Optimization with random Forests. The main objective of the proposed system is to improve the recognition rate and reduce the feature set size. The proposed system is trained and tested by a well-known classifier; Random forests (RF) on CENPRMI dataset. The proposed optimization algorithm obtained promising results in terms of classification accuracy as the proposed system is able to recognize 91.66 % of our test set correctly, as well as, it reduced the computational time. When comparing our results with other related works we find that our results is the highest among other published results.