Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Najdavan Abduljawad Kako has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Deep Learning for Cardiovascular Disease Detection

    by Shivan H. Hussein, Najdavan A. Kako

    Despite improvements, cardiovascular diseases (CVD) remain the most significant killer globally, accounting for around 17.9 million lives annually. Advancement of cardiac imaging modalities has taken place with Magnetic Resonance Imaging (MRI) along with artificial intelligence (AI) for changing scenarios of early diagnosis and management in cardiovascular diseases. This work investigates the role and contribution of deep learning, especially Fully Convolutional Networks (FCNs) and Convolutional Neural Networks (CNNs), toward the improvement of accuracy and automation in cardiac MRI analysis. The integration of AI enables accurate segmentation, efficient clinical workflows, and scalable solutions for resource-limited environments. A review of publicly available datasets underlines challenges in data variability and generalizability and points to the need for standardized models and explainable AI approaches. This work, therefore, underlines the possibility of improved diagnostic efficiency and equity in healthcare delivery using AI-driven methodologies in cardiovascular diagnostics. Future directions will focus on refining model scalability, enhancing dataset diversity, and validating clinical applications to foster robust and adaptable solutions.

  • Open Access

    Deep learning for early Parkinson's detection: A review of fundus imaging approaches

    by Zheen Ali, Najdavan Kako

    Parkinson's disease (PD), a type of neurodegenerative disease, is on the rise globally as the population ages. Today's costly diagnostic techniques for Parkinson's disease often detect the illness after significant brain damage has already occurred. Early detection is essential for improving patient outcomes and potentially slowing the disease's progression. One of the newest advances in artificial intelligence, deep learning (DL), presents new opportunities for the early, non-invasive diagnosis of Parkinson's disease. Fundus imaging, which captures fine-grained images of the retina, is a promising technique for detecting the disease's early symptoms. Changes in the retinal blood vessels and anomalies of the optic disc (OD) have been linked to neurodegeneration. DL models can identify subtle patterns in these fundus images, such as vascular alterations and changes in the optic disc, which have been connected to Parkinson's disease. This approach replaces current diagnostic methods with a scalable and cost-effective solution, increasing access to early detection. This review explores the current state of the art in using DL models with fundus images to detect PD early on, with a focus on significant public datasets, methodologies, and related research. It highlights how DL models could transform PD screening and provides an overview of the advancements and challenges in this emerging field.

  • An improved pelican optimization algorithm for function optimization and constrained engineering design problems

    by Haval Tariq Sadeeq, Araz Abrahim, Thamer Hameed, Najdavan Kako, Reber Mohammed, Dindar Ahmed

    Metaheuristic algorithms are a class of optimization techniques that have revolutionized problem-solving across various domains. These algorithms provide a versatile and powerful approach to finding near-optimal solutions for complex, combinatorial, and computationally intensive problems. They draw inspiration from natural processes, such as evolution, swarm behavior, or annealing, to iteratively refine solutions by intelligently navigating the problem space. Metaheuristics have become indispensable tools in both academia and industry, helping researchers and practitioners address real-world problems efficiently and effectively. The Pelican optimization algorithm (POA) is a recently developed metaheuristic algorithm that simulates the hunting behavior of pelicans. In complex optimization problems, an POA may have slow convergence or fall in sub-optimal regions, especially in high complex ones. In this paper, Levy flight is integrated into the exploration phase to enhance its search capabilities. Furthermore, a novel exponential parameter has been introduced to enhance the algorithm's overall performance by facilitating a smoother shift between exploration and exploitation phases. These modifications are intended to keep the algorithm from being locked in local optima. The developed algorithm named as IPOA was tested using widely recognized twenty-three benchmark functions with a variety of characteristics, a set of CEC2022 test suites, and five different engineering constrained problems. The results demonstrate the superiority and effectiveness of IPOA in tackling function optimization and constrained design engineering problems.


  • Open Access

    Multi-label deep learning for comprehensive optic nerve head segmentation through data of fundus images

    by Najdavan A. Kako, Adnan M. Abdulazeez, Diler N. Abdulqader


  • Peripapillary Atrophy Segmentation and Classification Methodologies for Glaucoma Image Detection: A Review

    by Najdavan A. Kako, Adnan Mohsin Abdulazeez

    Information-based image processing and computer vision methods are utilized in several healthcare organizations to diagnose diseases. The irregularities in the visual system are identified over fundus images with a fundus camera. Among ophthalmology diseases, glaucoma is the most common case leading to neurodegenerative illness. The unsuitable fluid pressure inside the eye within the visual system is described as the major cause of those diseases. Glaucoma has no symptoms in the early stages, and if it is not treated, it may result in total blindness. Diagnosing glaucoma at an early stage may prevent permanent blindness. Manual inspection of the human eye may be a solution, but it depends on the skills of the individuals involved. The diagnosis of glaucoma by applying a consolidation of computer vision, artificial intelligence, and image processing can aid in the prevention and detection of those diseases. In this review article, we aim to introduce numerous approaches based on peripapillary atrophy segmentation and classification that can detect these diseases, as well as details regarding the publicly available image benchmarks, datasets, and measurement of performance. The review article highlights the research carried out on numerous available study models that objectively diagnose glaucoma via peripapillary atrophy from the lowest level of feature extraction to the current direction based on deep learning. The advantages and disadvantages of each method are addressed in detail, and tabular descriptions are included to highlight the results of each category. Moreover, the frameworks of each approach and fundus image datasets are provided. Our study would help in providing possible future work directions to diagnose glaucoma.


  • Review on Image Segmentation Methods Using Deep Learning

    by Nabeel N. Ali, Najdavan A. Kako, Abdo Sulaiman Abdi

    In recent years, the machine learning field has been inundated with a variety of deep learning methods. Different deep learning model types, including recurrent neural networks (RNNs), convolutional neural networks (CNNs), adversarial neural networks (ANNs), and autoencoders, are successfully tackling challenging computer vision problems including image detection and segmentation in an unconstrained environment. Although image segmentation has received a lot of interest, there have been several new deep learning methods discovered with regard to object detection and recognition. An academic review of deep learning image segmentation methods is presented in this article. In this study, the major goal is to offer a sensible comprehension of the basic approaches that have already made a substantial contribution to the domain of image segmentation throughout the years. The article describes the existing state of image segmentation, and goes on to make the argument that deep learning has revolutionized this field. Afterwards, segmentation algorithms have been scientifically classified and optimized, each with their own special contribution. With a variety of informative narratives, the reader may be able to understand the internal workings of these processes more quickly.


  • Open Access

    DDLS: Distributed Deep Learning Systems: A Review

    by Et. Al. Najdavan Abduljawad Kako

    The clustered deep learning systems practice deep neural model networks with a cluster pooled resources aid. Distributed profound learning systems engineers should make multiple choices to process their diverse workloads successfully in their selected environment. Combined with the cluster bandwidth constraints, the abundance of GPU-based deep learning, the ever-greater size of data sets, and deep neural network models would entail developing high-quality models by distributed, profound learning systems designers. Because of their extensive lists of features and architectural deviations, it is not easy to compare distributed deep learning systems side by side. By examining the overall properties of deep learning models and how these workloads can be expanded into a cluster to carry out collective algorithm testing, the fundamental principles at work are shed when training a deep neural network in an isolated machinery cluster. Different techniques been addressed which are used by today's distributed deep learning systems and discuss their consequences. In order to conceptualize and compare deep-level structures, different methods have been developed by previous works to deep-level systems spread DDLS. Indeed, this paper addressed them to be more clearance for the readers.


  • Effect of Colored Noise on Neuron Membrane Size Using Stochastic Hodgkin-Huxley Equations

    by Najdavan A. Kako, Adnan M. Abdulazeez, Haval T. Sadeeq

    The investigations and experimentations show that ion channel noise in nerve cell can have deep effects on this cell behavior dynamically. More deeply, it has been seen that ion channel noise can rise stochastic resonance and spontaneous firing. In this paper, the effectiveness of colored noise on nerve cell membranes investigated under the relative standard deviation using stochastics Hodgkin-Huxley equations. The investigation detects that the colored noise term on small cell size consolidates the agreement with the microscopic numerical results. recently, it was found that a place is taken by non-trivial persistent cross correlation in a range, between the voltage of transmembrane fluctuations and the gate multiplicity. Experimentally, found that this phenomenon plays a main role in excitability rise and unprompted firing when the membrane cell become small. Also, the aforementioned phenomenon was found to increase the spike coherence remarkably. Moreover, Fox and Lu’s stochastic Hodgkin-Huxley equations and Güler model were expanded by combining colored noise in the conductance, to get some forms that ability to capture the addressed cross correlations. In the future, it is desirable to add white noise to the cell membrane and compare it with each of the colored noise, without color noise, microscopic, and Güler model to investigate the agreement between them in the membrane cell firing.

  • A New Hybrid Method for Global Optimization Based on the Bird Mating Optimizer and the Differential Evolution

    by Haval T. Sadeeq, Adnan M. Abdulazeez, Najdavan A. Kako, Dilovan A. Zebari, Diyar Q. Zeebaree

    Nature-inspired algorithms are often used by several diverse areas of engineering and science due to their easiness and versatility. Because metaheuristics operate by structurally changing and improving an established problem, they can often be extended to any optimization issues. The recent creation of metaheuristic algorithms has rendered them effective tools for solving NP problems. This paper presents a hybrid meta-heuristic method based on the Differential Evolution and Bird Mating Optimizer techniques to solve problems of global optimization. Bird Mating Optimizer is a novel method and is inspired by mating behavior of birds. Bird Mating Optimizer has some drawbacks such as producing poor results, trapping into local optima and slow convergence speed. Therefore, to conquer these insufficient it is hybridized with Differential Evolution approach. Differential Evolution technique is utilized to retain a preferable balance between both searches local and global. The performance and effectiveness of new Differential Evolution and Bird Mating Optimizer algorithm is tested and evaluated on 15 different functions of benchmark. The results of the experiment have shown the proposed technique possesses excellent performance in convergence speed, stability, and robustness, as compared to the well-known algorithms. It is proved that the Differential Evolution and Bird Mating Optimizer algorithm is very effective and superior to solve problems of global optimization. Experimental results indicate that the proposed hybrid Differential Evolution and Bird Mating Optimizer method is superior to previous existing state-of-the-art metaheuristic algorithms.


  • Open Access

    New symmetric key cipher capable of digraph to single letter conversion utilizing binary system

    by Najdavan Abduljawad Kako, Haval Tariq Sadeeq, Araz Rajab Abrahim

    In this paper, a new Playfair cipher built on bits level symmetric key cryptographic was proposed for the purpose of converting pairs of letters (digraphs) into single letters. The proposed algorithm is capable to overcome many of the shortcoming and vulnerabilities that exist in the current classical version of Playfair algorithm. The Playfair cipher is exceedingly complex than a classical substitution cipher, but still simple to hack using automated tactics. It is famous as a digraph cipher because two letters are exchanged by other two letters. This destroys any solo letter occurrence statistics, but the digraph statistics still unaffected (frequencies of two letters). Unluckily letter pairs have a flatter distribution than the one letter frequencies, so this intricacy matters for solving the code using pen and paper procedures. The suggested encryption process is conducted as follows; letters are first arranged in a spiral manner in Polybius square, afterwards, each pair will be replaced utilizing before-after technique if we are arranging pairs horizontally and down-up technique (vertically). The former process produces pairs of Plaintext that will be converted to binary bit stream then will be divided over blocks with stable sizes. Bits of these blocks are taken from pairs then fit them into square matrix of suitable order to put the concept of row-wise and revers row-wise matrix. Bits of this matrix are split into 2x2 square matrixes. The sub-matrixes are formed 8 bits. Here the XNOR operation is taken into consideration for bitwise operation to generate the keys for decryption and produce the cipher-text.