Journal of Kerbala University

Medicinal plants are considered vital resources of increasing importance at both lo-cal and global levels. tuse is not limited to developing countries; their significance extends to developed countries as well, where they are exploited as a primary source for extracting biologically active compounds used in the production of many modern medicines. These plants also contribute to supporting various sectors, including the food industry, by providing natural additives and healthy alternatives, as well as or-ganic farming, which relies on plant extracts for pest control and soil quality im-provement. Additionally, medicinal plants hold a prominent place in alternative and complementary medicine, as their popularity is increasing due to their connection to cultural heritage and the widespread belief in their relative safety compared to chem-ical treatments. Common examples of medicinal plants include mint, which is used to treat digestive disorders; chamomile, known for its calming and anti-inflammatory properties; ginger, which is used to relieve nausea and improve blood circulation; and turmeric, which contains compounds that are anti-inflammatory and antioxidant. Based on this, there is a need to enhance scientific studies and provide comprehen-sive and accurate information about these plants, including their chemical composi-tion, mechanisms of action, safe dosages, and potential interactions with medica-tions, ensuring their effective and safe use.
One essential natural resource that links environmental sustainability and hu-man health is medicinal plants. according to this study, using medicinal plants re-sponsibly not only enhances healthcare but also creates significant opportunities for biodiversity preservation and sustainable environmental and economic growth. Par-ticularly in areas where traditional medicine is the main form of treatment, medicinal plants can play a crucial role in integrated health systems through the adoption of sustainable agricultural practices, the implementation of legislative regulations, and the advancement of scientific research.
Rethinking how we use this botanical riches and making it a key component of com-prehensive sustain.

nic) are produced by two different types of fungi Aspergillus flavus and Aspergillus parasiticus. They are a worldwide concern for food supply, agricultural, and the pub-lic health economy, especially in tropical and sub-tropical regions around the world. The focus of this paper is to bring together recent research on types of aflatoxins, where they come from, and how they are harmful (affect people negatively) to hu-mans. Research on the food supply due to Climate Change, improper farming tech-niques, and improper food storage methods has resulted in an increase in the risk of aflatoxin contamination in foods worldwide. Long-term exposure to aflatoxins has been linked to causing liver cancer , weakening the immune system, stunting chil-dren's growth, and acute (sudden) poisoning. This notice will also discuss the new, sustainable ways to mitigate/prevent aflatoxin occurrences before or after harvest, such as using safe strains of bacteria/fungi to control fungus (toxigenic microbial biocontrol strains), having plants that are genetically resistant to fungi, using natural substances to prevent fungus (natural inhibitors), using new methods of storing food (hermetic storage technologies), using small particles for food safety (nanotechnolo-gy), detoxifying foods biologically (biological detoxification), and using agricultural methods that do not contribute to climate change (climate smart practices). Notwith-standing these advancements, there continue to be serious obstacles with respect to insufficient or nonexistent reporting infrastructure for surveillance/early identification systems; minimal understanding/education among small-holder farmers about this issue; and fragmented policymaking processes. The review concludes with a call for greater cooperation among various sectors, using an integrated One Health approach, to reduce the risks associated with aflatoxins and improve food security while safe-guarding the health of both people and the environment globally.

The study introduces 2,4,6-trichloroaniline as a novel coupling reagent for sensitive spectrophotometric determination of phenylephrine hydrochloride. The analytical method is based on the interaction of diazotized 2,4,6-trichloroaniline and phenylephrine hydrochloride under basic conditions to form a stable chromogenic compound. The compound exhibits strong yellow color and has a maximum absorbance wavelength at 410 nm, thereby enabling precise quantitative analysis. The method parameters such as temperature, reaction time, and reagent concentrations were carefully optimized to ensure improving sensitivity, high reproducibility and robustness of the procedure. Beer's law was found to be obeyed within the linear range of (0.5–12.0 μg/mL), limit of quantification (0.1289 μg/mL), limit of detection (0.0425 μg/mL), and (R² = 0.9997). The molar absorptivity coefficient was (1.58 × 104 L.mol-1·cm-1), which indicates the sensitivity of the method. Up to 60 minutes, the color of the product remained stable, and there was no significant change in the absorbance of the colored compound. For indicating the precision of the method, the results show the percentage of relative standard deviation of 0.16%, and an accuracy with an error percentage was 99.86% based on multi replicate analysis (n = 3). When the suggested method and the High-Performance Liquid Chromatography standard method were compared, the calculated t-value and F-test which were less than the critical t-value and F-value, indicating that there was no significant difference between the two methods. Overall, this method provides an efficient, cost-effective, and reliable means for the routine analysis of phenylephrine hydrochloride in pharmaceutical formulations.

Aims to shed more light on the potential role of plant-based phyto estrogens flax seed and soybean as dietary strategies to reduce the skeletal and hormo-nal alterations associated with poly cystic ovarian syndrome.Using female Albino Wistar rats, the experiment was carried out in the experimental ani-mal unit of a medical research facility.Research Design it was split up into eight groups: control group (G1) the standard casein protein diet was given .G2 is given soybean extract at a 100mg/kg concentration.G3 is given soy-bean extract at a dosage of 150mg/kg.G4: 200mg/kg soybean extract injec-tion Flax seed extract is administered to G5 at a 100mg/kg concentration. G6 consumes 150mg/kg flax seed extract for sustenance. G7Injection: 200mg/kg flax seed extract, G8:Synergistic between flax seed and soybean extract at 200 mg/kg .Poly cystic ovary syndrome was induced in female rats using a validated pharmacological model. The animals were administered letrozole at a dose of 1 mg/kg body weight daily for 21 consecutive days via oral administration to induce a hormonal disorder that mimics the characteristics of poly cystic ovary syndrome. The study results proved that the combined treatment with soy extract and flax seed at a con-centration of (200mg/kg) had a clear synergistic effect on sex hormones and the weights of internal organs in rabbits, compared to the G1. The results indicate that the combination of the two extracts may be more efficient than using each one individually, with no noticeable negative effects within the studied dosage limits, supporting the potential benefit of using them in fu-ture applied studies.

Background and Objective: Excessive use of chemical antiseptics like ethanol causes adverse skin conditions, while rising antimicrobial resistance necessitates sustainable al-ternatives. This study evaluated the antibacterial efficacy of aqueous leaf extracts of Ziziphus spina-christi (Sidr) from Karbala Governorate, Iraq, as an eco-friendly disin-fectant against pathogenic bacteria. Materials and Methods: Dried leaves were prepared using two aqueous extraction methods: infusion (40°C for 24 hours) and decoction (boiled to 100°C). Phytochemical screening tested for phenols, alkaloids, flavonoids, and sapo-nins. Antibacterial profiles were evaluated against five Gram-positive and four Gram-negative bacterial isolates using the agar well diffusion method and compared with Gen-tamycin (0.01 mg/ml). Results: Phytochemical screening confirmed an abundance of bio-active secondary metabolites including, alkaloids, flavonoids, phenols and saponin. Both aqueous preparations demonstrated significant broad-spectrum antibacterial activity. The aqueous infusion consistently exhibited superior inhibition over the boiled decoction across all strains. Infusion zones of inhibition ranged from 10±0.1 mm to 15±0.3 mm for Gram-positive bacteria, and 12±0.1 mm to 21±0.2 mm for Gram-negative bacte-ria. Escherichia coli was the most susceptible strain (21±0.2 mm), while Streptococcus pneumoniae was the least. Crucially, most tested clinical strains exhibited resistance to Gentamycin but remained highly vulnerable to the Sidr infusion extract. Conclusion: The aqueous infusion of Ziziphus spina-christi leaves possesses potent antimicrobial properties capable of overcoming conventional antibiotic resistance. Due to its energy-efficient prep-aration, lack of side effects, and robust antagonistic performance, it serves as a promising natural alternative for antiseptic and disinfectant formulations.

Background: Coronavirus syndrome is an acute contagious respiratory in-flammation caused by a strain of the coronavirus SARS-CoV-2, first found in China and then spread worldwide. Patients suffer from symptoms of varying severity, ranging from shortness of breath, low blood pressure, and oxygen deficiency to sepsis and vascular damage, which may lead to death. The ex-cessive inflammatory response caused by acute respiratory distress syndrome is the most common cause of death, and its pathogenic causes remain unclear to date. The erythrocyte sedimentation rate (ESR) is a laboratory test routinely used in clinical examinations as an indicator of acute-phase inflammation. ESR has been used in COVID-19 to assess disease severity and predict its course, aiming to study and explore its diagnostic role compared to several other response factors. Material and methods :This study included two groups of patients: the first group consisted of 55 patients infected with mod-erate severity cases, while the second group was composed of healthy individ-uals 29 as a control So it case control study. Results : were as follows: an in-crease in ESR level by p< 0.01 COVID-19 infected patients compared to healthy individuals, as well as an increase in Tumor Necrosis Factor Alpha (TNF-α) levels. The study also found no significant change in liver enzymes Alanine Transaminase (ALT)‎and Aspartate Aminotransferase (AST)between patients and the control group, while the relationship was found between ESR and CRP(0.325). Conclusion: The research concluded that the erythrocyte sedimentation rate (ESR) may reflect inflammatory activity in COVID-19 pa-tients

One of the most fascinating aspects of the merger between vision and language Artificial Intelligence is how to create human faces from textual data. This is one of the most challenging tasks in the field. The current state of the art in the field of text to facial generation is reviewed in this paper. From the first GANs to the most recent diffusions and transformers, we describe how and why the models use textual data to create faces, and what attributes and characteristics affect their output including informativeness of text, quality of the data-set, operational inefficiency, and dis/semantic gaps. We describe the bounding industry applications of the technology, including forensic facial reconstruction, security, entertainment, healthcare, and the arts, and we identify both potential and limitations of the technology. The technology is also applied to various fields and several ethics have been mentioned including the lack of human centered methodologies in assessing the data. We describe the most recent progress, the challenges faced in the field and the directions the field might take in the near future, and we present our efforts to ensure the ethical multi-faceted use of the technology. The review also describes in great detail the lack of ethical and evaluative methodologies in examining the potential of such systems. We present the review from this perspective in the hopes of providing the field with more rational and human use of such systems. This review distinguishes itself from prior works that focus on a single model or application. It provides an organized comparative summary of major text-to-face generation models — GANs, diffusion models, transformers — that have been assessed in the literature on dimensions of image fidelity, semantic coherence, efficiency, and ethics. It aims to help researchers pinpoint neglected areas in the field, understand model trade-offs, and formulate new avenues of research in this dynamic field.

An extensive review of real-time object identification methods for UAV-based applications is provided in this work. The study focuses on mul-timodal frameworks, lightweight detection techniques, small-object detec-tion methods, and deep learning-based approaches, especially YOLO-based models. The annotated publications access methodology for the stated peri-od is reminiscent of studies published from 2019 to 2025. The publications were sourced through the use of relevant keywords associated with UAV object detection and real-time systems within the vast scientific databases of IEEE Xplore, ScienceDirect, and SpringerLink.
The selected research papers were analyzed and classified into four main categories: multimodal detection systems, lightweight and edge computing models, attention mechanisms for small object detection, and YOLO-based detectors. Research papers were compared and evaluated based on detection accuracy, computational efficiency, and real-time operational viability on resource-constrained UAV platforms. The results indicated that YO-LO-based models, while achieving rapid inference times, were still unable to detect small and spatially clustered objects sufficiently. On the other hand, advanced designs tended to significantly increase the computational burden leading to negative impact on the real-time aspect of the system, even though the designs improved detection accuracy.
The assessment identifies and highlights potential research directions, in-cluding practical design solutions for lightweight models and the challenges posed by small object size, limited computational resources, and the need for real-time performance.

Facial Emotion Recognition enables machines to detect human af-fect and react appropriately. Standard transformer FER networks demand heavy computation - they fail on low power hardware that must work in real time. We therefore introduce MobileViT-SECA, a small hybrid net-work that extends the MobileViT backbone with two attention blocks arranged in series - first a Squeeze-and-Excitation (SE) layer then a Co-ordinate Attention (CA) layer. The SE→CA stack raises channel selec-tivity and sharpens spatial focus but keeps parameter count low. Trained on the eight class AffectNet corpus, the network reaches an F1-score of 86.2 % and retains stable results on further FER test sets. Ablation tests verify that the ordered attention pair yields clear gains over the standalone backbone. Gradient based visual inspection reveals that the network locks onto salient facial areas before it outputs a label. Mo-bileViT-SECA thus trades a minor accuracy drop for a major cut in compute load - it fits emotion aware software that must run on everyday edge hardware

Community detection plays a central role in analyzing complex real-world networks such as social and collaboration graphs. However, in many practical scenarios, communities are overlapping, meaning that a node may belong to multiple groups simultaneously. Recent deep learning approaches, notably Neural Overlapping Community Detection (NOCD), have addressed this challenge by combining probabilistic modeling with graph neural networks, while Attention-based Overlapping Community Detection (AOCD) further improves representation learning through graph attention mechanisms. De-spite these advances, existing methods typically rely on a fixed graph topolo-gy during training, which can limit their ability to handle noisy inter-community links in overlapping settings. In this work, a community-aware DropEdge framework has been proposed, built on top of an AOCD-style graph attention encoder. The proposed method adaptively removes edges with low feature coherency, measured using cosine similarity between normalized node features, while preserving full context adjacency for evaluation. This design reduces structural noise during training and strengthens the communi-ty-consistent message without changing the probabilistic purpose of the origi-nal model. Experimental results on benchmark datasets with overlapping ground truth communities show that the proposed approach consistently out-performs the baseline in all 10 evaluated datasets. On average, this improves retrieval performance by 1.2 NMI points (about 3.35% relative gain), with the largest improvement reaching 4.2 NMI points. These findings indicate that community-aware structural regularization can be effective.

Despite the considerable progress made in speaker recognition technologies, maintaining robust performance under real-world noisy recording conditions remains a challenging task, as environmental noise can significantly degrade system accuracy. This work focuses on a hybrid deep learning architecture that integrates convolutional neural networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM)layers to improve speaker recognition in noisy environments. Speech signals sampled at 16 kHz are pre-processed using zero padding to ensure uniform input length, followed by feature extraction in the frequency domain. The CNN layers are used to learn distinct spectral features, and the BiLSTM layers are used to learn temporal dependencies. The pro-posed CNN–BiLSTM architecture was evaluated on the PCM16k dataset in noisy environments with added Gaussian white noise (AWGN) at various sig-nal-to-noise ratios (SNR). Background noise was also added, and the CMU Arctic dataset was used to test the generalization capability of the proposed model. The experimental results show that the proposed CNN–BiLSTM model outperforms the CNN–LSTM model, achieving a high accuracy of 99.60% and maintaining robust performance even at SNR = 0 dB, where the model achieves an accuracy of 86.42%. These findings demonstrate the effectiveness of the proposed CNN–BiLSTM architecture in enhancing speaker identifica-tion performance and improving robustness against severe noise conditions.

Deep learning has brought about a change in the ways images are hid-den, allowing the transition from manual embedding techniques to traina-ble, data-driven models. This review presents a comprehensive and struc-tured study of deep learning-based image steganography techniques , cate-gorizing current approaches into five common architectural categories: convolutional neural networks (CNNs), Generative Adversarial Networks (GANs) , transformer-based models, diffusion-based generative architec-tures, and autoencoder-based architectures. This review analyzes and com-pares the architectures, embedding methods, performance characteristics, and evolutionary trends of existing models , with a focus on undetectable capabilities, payload capacity, flexibility, security against steganalysis attacks , and ease of deployment. The review begins with early CNN-based encoder–decoder models and extends to adversarial, attention-based, se-mantic, probabilistic, and reconstruction-driven steganographic frame-works. It further examines how representation learning, self-attention mechanisms, adversarial optimization, and stochastic generative modeling have influenced the development of modern image steganography systems. In addition to synthesizing recent advances, this study identifies persistent research gaps, including limited cross-domain generalization, insufficient explainability, computational complexity, inconsistent robustness evalua-tion, and the absence of standardized benchmarks. By unifying develop-ments across different architectural families and generations, this review provides a focused taxonomy and a critical understanding of the current state of deep learning-based image steganography. It also outlines future research directions toward more adaptive, explainable, robust, secure, and practically deployable steganographic systems.

The present-day uses of the Internet as a platform for e-commerce have broadened the avenues for e-business. With e-commerce platforms gen-erating myriad online reviews, end users, when making online purchas-es, rely on these reviews, and for businesses, the integrity of these re-views is of utmost importance as it can impact their bottom line. For such systems, reliable methods for identifying genuine reviews from false ones pose an important problem. With this in mind, the present study delves into a number of systems that have been developed to de-tect fraudulent reviews. The problem of review fraud is compounded by the increasing sophistication of fraudulent writing, the scarcity of relia-ble datasets with verified labelling, and the linguistic diversity between genuine and fraudulent reviews. This study endeavours to build a survey to cover the existing, to the best of our knowledge, anomaly detection systems aimed at online user reviews, from statistical, machine learning (ML), deep learning (DL), and transformer-based systems, along with their strengths and weaknesses.

The Arbaeen pilgrimage is one of the largest annual mass gatherings in Iraq, posing major challenges for health services due to high participant density and limited infrastructure, which may increase disease transmission. This article presents a digital design supported by a data governance framework for rapid response to serious health conditions during the pilgrimage. The electronic system uses Google Forms to systematically collect patient data, including age, gender, symptom onset location, therapy type, and clinical outcome. The system automatically converts this data into real-time read-ings displayed on a central monitoring screen. The paper discusses the tech-nical architecture, data-quality assurance measures, early warning indicators for potential outbreaks, implementation, training, and risk management strategy. The electronic system is a low-cost, responsive technology de-signed to reduce response time and facilitate coordination among hospitals and medical units, with data as a key factor in delivering fast, effective health care during such events, thereby reducing morbidity and mortality.

Background: Liquid biopsy of circulating tumor DNA (ctDNA) is a groundbreaking technology that facilitates cancer screening and surveillance on a much less invasive basis than conventional tissue biopsies. Aims: The literature presented in this review integrates the clinical applications of ctDNA for early diagnosis, treatment response, minimal residual disease (MRD) followed up longitudinally, and resistance mutation characterization.
Methodology: A literature review was conducted in global biomedical databases across peer-reviewed studies of polymerase chain reaction (PCR) and next-generation sequencing (NGS) methods in liquid biopsy processes. Literature Review: Based on all literature up to October 2023, the current state of literature related to polymerase chain reaction (PCR) in liquid biopsy workflows for this topic is extensive. The liter-ature realizes that PCR and NGS techniques have equally improved oncology, but common application suffers from formidable limitations. This includes gaps in ge-nome sequence amplification and biological interference through clonal hematopoie-sis of indeterminate potential (CHIP) and overpriced methods.
Results: The data suggest that new technologies especially artificial intelligence, advanced bioinformatics, and multi-analytic biomarkers (methylation patterns, mi-croRNAs, and proteins) are addressing some of the technological challenges. Moreo-ver a recent study emphasizes the applicability of liquid biopsy principles to neuro-logical and cardiologic fields.
Conclusion: Liquid biopsy in precision medicine will inevitably depend on extensive laboratory assay standardization, cost-control, and sound clinical validation. Over time, as these techniques mature, liquid biopsy will transform from being a classical cancer management tool to a more personalized therapeutic method for people of dif-fering disease profiles, transforming to a state of clinical workflows oriented toward the specific client.