مواقع التدريسيينجامعة الكوفة
0Publications
0Lectures
0Interests
0News

About

Have a Bachelor's degree in computer science from the University of Al-Qadisiyah, in 2009.
And a Master's Degree in Computer Science Department of Artificial Intelligence and Robotics from the Ferdowsi University of Mashhad, in 2017.
And a Ph.D Degree in Computer Science Department of Artificial Intelligence and Robotics from the Ferdowsi University of Mashhad, in 2023.

Specialization: Artificial Intelligence.
Current position: head of the Information Technology Division at the College of Veterinary Medicine.
Current place of work: University of Kufa - College of Veterinary Medicine.

Research interests in the field of Artificial Intelligence and Robots, Image Processing, and Machine Learning.

Courses and experiences:
1- A course to develop and qualify the capabilities of university leaders and human resources.
2- Managing the college website.
3- Managing the Moodle E-learning system.
4- Management and maintenance of computer laboratories in the college.
5- Managing E-mail for all college members.

Research Interests

6

Publications

5
2026

An end-to-end distribution- and task-aware parallel imputation method for improving target task performance

AuthorsKarrar Al-Kaabi, Davood Zabihzadeh
JournalEngineering Research Express
Date2026-08-12
Abstract

Abstract Most machine-learning imputation techniques treat missing values independently of the downstream task, resulting in suboptimal predictive performance. While some recent methods jointly train an imputer with a target predictor, they fail to produce diverse and context-sensitive imputations and suffer from training inefficiencies. Additionally, in many methods, missing values are ignored during training, and the imputers are trained only on observed data. To overcome these limitations, we propose an imputation method, named end-to-end task-aware parallel Imputation with class-wise prototypes (ETPI). It captures the class-conditional distributions of input data using a few proxies. For the missing entries in any sample, ETPI generates class-aware pseudo-targets based on the predictor loss and the estimated class-conditional distribution. By exposing the imputer to both observed data and optimal pseudo-targets during training, ETPI effectively leverages training information to fit the imputer model and aligns imputation with the objectives of the target task. Extensive experiments on classification and regression tasks show that ETPI outperforms other state-of-the-art methods. It also maintains high imputation quality even with limited training data or high missing rates, mainly due to the high quality of the generated pseudo-targets and the integration of imputation and prediction into a single end-to-end pipeline.

2026

IntelliShard++: A Hybrid Reinforcement Learning and Particle Swarm Optimization Framework for Dynamic Blockchain Sharding

AuthorsSajjad H. Hasan, Karrar Al-Kaabi Firas H. Neamah Hussein Alrammahi
JournalHighTech and Innovation Journal
Date2026-09-01
Abstract

Abstract Sharding lifts blockchain throughput by splitting validators into parallel committees, but deciding how many shards to run, at what capacity, and when to change them is still unsettled. The objective of this study is to control shard count and pershard capacity jointly under an explicit reconfiguration cost. IntelliShard++ pairs a tabular epsilon-greedy Q-learning agent, which decides when to act, with an Adaptive Guided Best-Worst Particle Swarm Optimizer, which chooses the shard count and capacities whenever a rebalance is triggered. Both components share an objective that charges for restructuring. The framework was evaluated on an Ethereum-calibrated workload of 100,000 windows against a ten-shard baseline, a twenty-shard baseline, and a reactive threshold controller tuned with cooldown and hysteresis. IntelliShard++ served 2,290 transactions per window, which is 45.7% above the ten-shard baseline and 9.9% above the twenty-shard baseline, at the lowest mean latency of any system tested and with finality shortened by 18.5%. Against the tuned reactive controller, the margin was 9.1%. On five unseen workload seeds, a frozen policy gained 16.4% over the strongest static baseline, so the learned policy transfers. Worst-case latency and Byzantine margin degrade, and both are reported.

2017

A Review of Lung Nodule detection and segmentation in Chest X-ray radiographs

Authorsكرار علي حسين, مهدي سعادتمند طرزجان
Journalالجمهورية الاسلامية الايرانية جامعة صنعتي شريف
Abstract

Abstract Introduction: Lung cancer is one of the deadliest type of cancer for both men and women, all around the world, which is the highest mortality rates in all cancers. Hence early detection of lung cancer is the most promising way to improve the patient's chances of survival, and is one of the best possible clinical outcomes for patients. Chest X-ray (CXR) is the most common test used as a primary diagnostic tool for a variety of clinical conditions, which includes at least one-third of all radiology tests. Many diseases, such as lung cancer, can be diagnosed in the early stages by various approaches such as periodic health screenings. Lung cancer can appear by observing a single nodule in CXR images, thus recognizing pulmonary nodules can provide a significant impact on early detection of lung cancer. Method analysis: Chest X-ray is an effective tool for detecting lung tumors, but interpretation of these images are difficult due to the anatomical structure of the image. An experienced radiologist sometimes faces a disturbance in the X-ray image. Different computer-assisted diagnostic methods have been proposed to help the radiologist find disturbances of x-ray images, but none of them have a complete effect due to the complexity of the anatomical structure in the image and the tiny nodules. In this study, various methods for finding a tumor from chest radiographs from document databases around the world have been investigated. These methods often are tested on standard images databases, such as the famous JSRT database. Most methods in articles are categorized in some stages: preprocessing, main processing involves lung field detection, feature extraction (with different methods) to find candidate nodules, and then reducing the false nodule candidates (FP). Finally, a classification method estimate real nodules. A brief description of each of these methods is given and listed in a table in full manuscript. Conclusion: Large areas of Chest X-ray images, like bones and organs in the image, also excessive amount of noise and tissue-to-tissue interactions with the lung and its tumors are shown in CXR image. The studies have shown that these challenges, in addition lack of in-depth methods, cause false nodules to be detected, thereby reducing evaluation criteria values. Various methods have been proposed and reported various criteria such as accuracy, precision, specificity, sensitivity, area under curve (AUC) and so on. Nowadays, deep neural network approach have been used to solve this problem, but all of the presented methods have their own unique advantages and disadvantages, and none of them has complete results due to the challenges posed and tiny nodules. We wish that new proposed methods can be used to solve this problem using the high abilities. Keywords: computer-aided diagnosis (CAD), chest X-ray image, lung nodule, feature extraction, classification.

2020

CCBlock: an effective use of deep learning for automatic diagnosis of COVID-19 using X-ray images

AuthorsAli Al-Bawi, Karrar Al-Kaabi Mohammed Jeryo Ahmad Al-Fatlawi
JournalResearch on Biomedical Engineering
Date2020-10-30
Abstract

Propose Troubling countries one after another, the COVID-19 pandemic has dramatically affected the health and well-being of the world’s population. The disease may continue to persist more extensively due to the increasing number of new cases daily, the rapid spread of the virus, and delay in the PCR analysis results. Therefore, it is necessary to consider developing assistive methods for detecting and diagnosing the COVID-19 to eradicate the spread of the novel coronavirus among people. Based on convolutional neural networks (CNNs), automated detection systems have shown promising results of diagnosing patients with the COVID-19 through radiography; thus, they are introduced as a workable solution to the COVID-19 diagnosis. Materials and methods Based on the enhancement of the classical visual geometry group (VGG) network with the convolutional COVID block (CCBlock), an efficient screening model was proposed in this study to diagnose and distinguish patients with the COVID-19 from those with pneumonia and the healthy people through radiography. The model testing dataset included 1828 Xray images available on public platforms. Three hundred and ten images were showing confirmed COVID-19 cases, 864 images indicating pneumonia cases, and 654 images showing healthy people. Results According to the test results, enhancing the classical VGG network with radiography provided the highest diagnosis performance and overall accuracy of 98.52% for two classes as well as accuracy of 95.34% for three classes. Conclusions According to the results, using the enhancedVGG deep neural network can help radiologists automatically diagnose the COVID-19 through radiography.

2022

A framework to enhance generalization of deep metric learning methods using general discriminative feature learning and class adversarial neural networks

AuthorsKarrar Al-Kaabi, Reza Monsefi Davood Zabihzadeh
JournalApplied Intelligence
Date2022-08-02
Abstract

Deep Metric Learning (DML) methods automatically extract features from data and learn a non-linear transformation from the input to a semantically embedding space. Many DML methods focused to enhance the discrimination power of the learned metric by proposing novel sampling strategies or loss functions. This approach is very helpful when both the training and test examples are selected from the same set of categories. However, it is less effective in many applications of DML such as image retrieval and person-reidentification. Here, the DML should learn general semantic concepts from observed classes and employ them to rank or identify objects from unseen categories. Neglecting the generalization ability of the learned representation and just emphasizing to learn a more discriminative embedding on the observed classes may lead to the overfitting problem. To address this limitation, we propose a framework to enhance the generalization power of existing DML methods in a Zero-Shot Learning (ZSL) setting by general yet discriminative representation learning and employing a class adversarial neural network. To learn a general representation, we employ feature maps of intermediate layers in a deep neural network and enhance their discrimination power through an attention mechanism. Besides, a class adversarial network is utilized to force the deep model to seek class invariant features. We evaluate our work on widely used machine vision datasets in a ZSL setting. Extensive experimental results confirm that our framework can improve the generalization of existing DML methods, and it consistently outperforms baseline DML algorithms on unseen classes.

Lectures

7

مكونات الحاسوب

الحاسوب

نظام التشغيل

الحاسوب

تطبيقات الحاسوب- Microsoft Word

الحاسوب

تطبيقات الحاسوب- Microsoft PowerPoint

الحاسوب

تطبيقات الحاسوب- Microsoft Excel

الحاسوب

تطبيقات الحاسوب- Microsoft Access

الحاسوب

الذكاء الإصطناعي - Artificial Intelligence

الحاسوب

News & Announcements

7
News 2026-11-14

Workshop on Cybersecurity and Personal Data Protection

As part of its scientific and awareness-raising activities, a workshop entitled "Cybersecurity and Personal Data Protection" was held, presented by Lecturer Dr. Karrar Ali Hussein Al-Kaabi. The workshop addressed the concept of cybersecurity and its importance in protecting digital systems and information, highlighting the major cyber risks and threats that users may encounter and the measures required to prevent them. It also presented a set of guidelines and safe practices for using the Internet and digital platforms, particularly in relation to protecting online accounts, creating secure passwords, handling suspicious links and messages with caution, and maintaining the privacy of personal data. The workshop aimed to raise awareness of digital security and promote a culture of safe and responsible technology use, thereby contributing to the protection of personal information and the reduction of cybersecurity risks within the university environment and the wider community.

News 2026-11-14

Workshop on Artificial Intelligence in Higher Education

As part of its scientific and academic development activities, a workshop entitled "Artificial Intelligence in Higher Education" was held, presented by Lecturer Dr. Karrar Ali Hussein Al-Kaabi. The workshop introduced the concept of artificial intelligence and its major applications in higher education, highlighting its role in developing teaching and learning methods and enhancing the efficiency of the educational process. It also explored a range of AI-powered tools and applications that can be utilized in developing educational content, improving assessment methods, supporting scientific research, facilitating students' access to information, and strengthening their self-directed learning skills. Furthermore, the workshop emphasized the importance of the responsible and ethical use of artificial intelligence technologies, highlighting the need to maintain scientific accuracy and academic integrity when incorporating these technologies into the university environment. The workshop aimed to raise awareness among faculty members and students of the potential of artificial intelligence and encourage the effective and responsible integration of its applications into higher education, thereby keeping pace with technological advancements and contributing to the improvement of university education.

News 2025-04-28

Workshop on E-Learning Applications and Their Role in Veterinary Medicine Education

As part of its scientific activities, the Faculty of Veterinary Medicine organized a workshop entitled "Applications of E-Learning in Veterinary Medicine Education," presented by Assistant Professor Mohammed Taha Naqi Al-Bayati and Lecturer Dr. Karrar Ali Hussein Al-Kaabi, affiliated with the Ibn Sina Center for E-Learning at the Faculty of Veterinary Medicine, University of Kufa. The workshop discussed websites and electronic platforms that support and enhance the educational process in a manner tailored to the teaching of scientific subjects and the practical nature of veterinary education. The speakers also explained how to integrate these tools with the Moodle platform used at the University, aiming to strengthen students' skills and improve academic outcomes.

News 2024-01-31

Holding a workshop on how to raise the evaluation of the college’s website

The Information Technology Division at the Faculty of Veterinary Medicine held a workshop entitled “Mechanisms for Raising the Faculty Website Evaluation” on Monday, 1/2/2024, delivered by the Lecturer, Dr. Karar Ali Al-Kaabi, head of the Information Technology Division at the Faculty. The lecturer touched on the importance of following the correct mechanisms and conditions for publishing various news on the college’s website and diagnosing weak points. He also stressed the need to take care of the quality of images and according to the publishing controls for the websites of the colleges of the University of Kufa to improve the evaluation of the college’s website.

News 2024-01-31

Holding a Workshop entitled Drafting News and the Mechanism of Publishing it on the College Website

The Computers and Internet Division at the Faculty of Veterinary Medicine held a workshop entitled (Formatting news and the mechanism for publishing it on university websites), in which Dr. Karrar Ali Hussein Al-Kaabi, Responsible for the Computers and Internet Division, lectured. The workshop included several topics, the most important of which was how to formulate news about scientific activities, urging the teaching staff on the importance of the media aspect, a course in raising the level of evaluation of the college’s website, and the mechanism for selecting photographs. The workshop also covered a review of the recent evaluation from the Information Research and Rehabilitation Center for the college’s website, presenting all its paragraphs, and ways to avoid previous mistakes and processed.

News 2018-03-28

Attend the Master's discussion

Attendance of the discussion of the master student Ahmed Abdul Razzaq Al-Fatlawi, which was held at Al-Kafil National University in cooperation with the Ferdowsi University of Mashhad

News 2018-03-28

Participation in the cultural day of the college

Participation in the cultural day of the college where the ceremony was started with the word of the Dean and the Assistant Administrative Assistant and then the presentation of a student play and the dumping of national poetry and the participation of students in the photo gallery.

View details