CTU Journal of Innovation and Sustainable Development (CTUJoISD), ISSN 2588-1418 and e-ISSN 2815-6412, is an international double-blind peer-reviewed journal that publishes original and high-quality research articles and review articles in multi-disciplines. It previously existed as Can Tho University Journal of Science (CTUJS) which was established in 2015 with assigned codes ISSN 2615-9422 and e-ISSN 2815-5602.

The aim of the Journal is to be a key source of research articles to provide scientific research results of Can Tho University in particular, and domestic and foreign scientific research projects in general, contributing to promoting scientific research and technology transfer.

The scopes of the Journal include, but not limited to, the following topic areas:
1. Agricultural and Biological Sciences;
2. Business, Management and Accounting;
3. Engineering;
4. Social Sciences;
5. Multidisciplinary.

Editor-in-Chief
Tran Ngoc Hai
Professor
Can Tho University, Viet Nam
Research Interests: Advances in Seed Production for Aquaculture, Advances in Aquaculture, Aquaculture Planning and Development, Integrated Coastal Zone Management
 
Deputy Editor-in-Chief
Tran Thanh Dien
PhD
Can Tho University, Viet Nam
Research Interests: Recommender Systems, Data mining in education

Editorial Board Members

Tetsu Ando
Professor
Tokyo University of Agriculture and Technology, Japan
Research Interests: Molecular Mechanism of Bio-Interaction

Fu-Sung Chiang
Professor
National Taiwan Ocean University, Taiwan
Research Interests: Consumer Economics, Fisheries/Aquaculture Economics, Demand and Market Analyses, Marketing and Trade

Nguyen Ngoc Dien
Associate Professor
Hoa Sen University, Viet Nam
Research Interests: Civil Law

Nigel K. Downes
PhD
GIZ/CIM Integrated Expert
Research Interests: Environmental Science, Geography

Phan Trung Hien
Associate Professor
Can Tho University, Viet Nam
Research Interests: Administrative Law

Le Quoc Hoi
Professor
National Economics University, Viet Nam
Research Interests: Economic growth, FDI, poverty and income inequality

Atsushi Ishimatsu
Professor
Nagasaki University, Japan
Research Interests: Environmental Physiology, Comparative Physiology, and Morphology

Samir Kumar Khanal
Professor
University of Hawaii at Manoa, USA
Research Interests: Anaerobic Digestion, Aquaponics, Waste-to-Resources, Environmental Biotechnology

Nguyen Dac Khoa
Associate Professor
Can Tho University, Viet Nam
Research Interests: Rice, Plant Biotechnology, Crop Management

Phan Dinh Khoi
Associate Professor
Can Tho University, Viet Nam
Research Interests: Microfinance, Microeconomic Theory, Behavioral Finance

Nguyen Ngoc Lam
Professor
Institute of Oceanography, Viet Nam
Research Interests: Marine Biology, Phytoplankton, Harmful Algal Blooms, and Dinoflagellates

Juan Boo Liang
Professor
Universiti Putra Malaysia, Malaysia
Research Interests: Animal Nutrition, Livestock Waste Management

Hoang Ngoc Long
Professor
Institute of Physics, Viet Nam Academy of Science and Technology, Viet Nam
Research Interests: Field and particle theory

Juan J. Loor
Professor
Department of Animal Sciences, University of Illinois, United States
Research Interests: Nutrition, Physiology, Genomics, Lactation, Dairy, cow

Do Thanh Nghi
Associate Professor
Can Tho University, Viet Nam
Research Interests: Mining Complex Data, Support Vector Machines, Decision Trees, Ensemble-based Learning, Information Visualization

Nguyen Chi Ngon
Associate Professor
Can Tho University, Viet Nam
Research Interests: Intelligent Control

Nguyen Trong Ngu
Associate Professor
Can Tho University, Viet Nam
Research Interests: Animal and Veterinary Sciences, Animal Breeding, Animal Husbandry, Veterinary Medicine

Minh Nguyen
PhD
The Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia
Research Interests: Climate Adaptation and Sustainable Development

Pham Thi Hong Nhung
Associate Professor
University of Foreign Languages of Hue University, Viet Nam
Research Interests: Language Education, Intercultural Communication, Pragmatics

Kazufumi Osako
Professor
Tokyo University of Marine Science and Technology, Japan
Research Interests: Life sciences, Aquatic Life Sciences, Food Sciences

Ngo Thanh Phong
Associate Professor
Can Tho University, Viet Nam
Research Interests: Chemistry, Organic Synthesis

Nguyen Thanh Phuong
Professor
Can Tho University, Viet Nam
Research Interests: Adaptation of aquatic animals to environmental factors, Sustainable production of aquaculture systems, Hi-tech aquaculture production systems

Yeong Yik Sung
Professor
Universiti Malaysia Terengganu, Malaysia
Research Interests: Stress proteins, Aquaculture

Yuji Tanaka
Professor
Tokyo University of Marine Science and Technology, Japan
Research Interests: Plankton Oceanography

Nguyen Thanh Thuy
Professor
VNU University of Engineering and Technology, Viet Nam
Research Interests: Artificial Intelligence, Soft Computing, Hybrid Intelligence, Knowledge-Based Systems, High Performance Computing, Grid Computing

Tran Trung Tinh
Associate Professor
Can Tho University, Viet Nam
Research Interests: Electrical Systems, Transmission System, Smart Grid

Nguyen Hieu Trung
Associate Professor
Can Tho University, Viet Nam
Research Interests: Water Management, Land Use Planning

Phuong Hoang Yen
Associate Professor
Can Tho University, Viet Nam
Research Interests: English Language Teaching, Student Learning Autonomy, Teacher Professional Development, Testing and Assessment in Language Teaching

 

Vol. 18 No. 1 (2026)

Published: 2026-03-24

Biological Activity of Extracts from OM5930 Rice Components in Controlling Weedy Rice under Laboratory Conditions

Ho Le Thi, Nguyen Gia Huy, Vu Nhat Vy, Le Nha Tran, Kieu Cong Vinh, Nguyen The Cuong
Abstract | PDF
Weedy rice (Oryza sativa f. spontanea) poses significant challenges to rice production, reducing yield and commercial value. This study investigates the inhibitory effects of root, stem, and leaf extracts from cultivated rice OM5930 (60 days after sowing) on two weedy rice lines, WR19 (short awn) and WR20 (long awn). The objectives were to: (i) identify the most potent plant part for inhibition, (ii) determine the optimal treatment duration, and (iii) assess the resistance of the two weedy rice lines. Results demonstrated that extracts from all OM5930 plant parts suppressed seedling shoot and root growth in both weedy rice lines. Leaf extracts exhibited the strongest inhibition, achieving complete suppression (100%) at 0.3 g/mL across all time points (0, 48, and 96 hours). The 48-hour treatment showed the most stable inhibitory effect. Root and stem extracts displayed lower efficacy, reaching only 60–70% inhibition at the same concentration. WR20 was more susceptible than WR19, with a slight stimulation observed at lower concentrations (0.015 and 0.075 g/mL). The strong allelopathic potential of OM5930 leaf extracts suggests their application as an eco-friendly bioherbicide, reducing dependence on synthetic herbicides, minimizing production costs, and promoting sustainable rice cultivation.

Effects of collection frequency and age on semen characteristics of crossbred buck rabbit (New Zealand White × Local)

Truong Thanh Trung, Tran Thi Thanh Khuong, Tran Long Hai
Abstract | PDF
The study involved 24 crossbred bucks (New Zealand White × Local) in a factorial design to assess the impact of semen collection frequency and age on semen quality. Two factors were considered: the age of the bucks (5-6 months for young and 9-10 months for mature) and the frequency of semen collection (daily (Fre-1), every three days (Fre-3), every five days (Fre-5), and weekly (Fre-7)). The study was conducted over 12 weeks. Results from the last 7 weeks indicated that weekly semen collection yielded the best outcomes, with a concentration of 346×10⁶ sperm/mL, 66.7% sperm motility, 70.5% membrane integrity, and 66.8% live sperm rate. Furthermore, semen volume and the number of motile sperm increased as the collection interval increased, with mature bucks (9-10 months) showing better semen characteristics than younger bucks. Younger bucks had smaller testicular dimensions than mature bucks. In conclusion, the optimal semen collection frequency for New Zealand White × Local crossbred bucks was once a week, and the recommended age for semen collection was 9-10 months.

Effect of probiotics and vitamins supplements on reproductive and egg quality of Ac chicken at 40-50 weeks old

Le Thanh Phuong, Pham Tan Nha
Abstract | PDF
The study investigated the impact of probiotics and vitamins (PV) on the reproductive performance and egg quality of Ac chickens' diet. The experiment was conducted using a completely random design with 250 hens, aged 40-50 weeks, housed in cages. There were 5 treatments, each with 5 repetitions, and 10 hens per repetition. The treatments corresponding to the diets were: Control (BD), C250 (BD + 250 mg PV/kg of feed), C500 (BD + 500 mg PV/kg of feed), C750 (BD + 750 mg PV/kg of feed), and C1000 (BD + 1000 mg PV/kg of feed). The results showed that the highest laying rate (39.2%), yolk index (0.45), albumen index (0.08) and yolk colour (7.6) were in C1000. Although the C750 indicated the lowest FCR (3.78), it had the greatest egg weight (36.4 g/egg) and Haugh unit (82.8). In conclusion, probiotics and vitamins at 1,000 mg/kg diet improved Ac chicken laying rate, FCR, and egg weight.

Isolation of sodium benzoate-degrading bacteria from rice noodle manufacturing wastewater in Can Tho, Viet Nam

Nguyen Tran My Han, Pham Nhu Huynh, Vo Phat Tai, Nguyen Thi Hong Cam, Tran Ngoc Que Linh, Nguyen Thi Phi Oanh
Abstract | PDF
Diverse starch-based foods have been widely consumed. To keep the products in good quality, preservatives are added. Sodium benzoate is one of the most commonly used food preservatives for starch-based products. As an antimicrobial compound, high concentrations of sodium benzoate in water can pose risks to aquatic indigenous microbial communities. From wastewater samples of three rice noodle facilities, 27 sodium benzoate-degrading bacteria were isolated on minimal mineral medium supplemented with sodium benzoate (50 mg/L). Isolates SB2.1, SB2.2, SB2.4, SB3.10, and SB3.13 demonstrated their best sodium benzoate degradation efficacy in liquid minimal mineral medium containing 1,250, 2,500, and 5,000 mg/L sodium benzoate, quantitatively measured by the spectrophotometric method at 225 nm. Optimal degradation efficacy (>93%) at 1,250 mg/L sodium benzoate was observed when isolates SB2.1, SB2.2, SB2.4, and SB3.10 were grown in a minimal mineral medium at pH 7-9. Notably, SB3.10 exhibited chemotaxis towards sodium benzoate at 24 hours of incubation. Based on 16S-rRNA gene sequencing and biochemical characterization, SB3.10 was identified as Acinetobacter calcoaceticus SB3.10.

A comparative deep learning approach for image classification and retrieval in scientific publications

Nguyen Hoang Anh, Tran Thanh Dien
Abstract | PDF
This study presents a comparative analysis of state-of-the-art deep learning models–EfficientNetB0, MobileNetV2, and ResNet101–for image classification and content-based retrieval in scientific publications. A dataset of 4,303 images from 11 categories was curated from the Can Tho University Journal of Science and enhanced through tailored data augmentation strategies. The models were fine-tuned using transfer learning with hyperparameters optimised via Grid Search. Features were extracted using GlobalAveragePooling2D, and cosine similarity was combined with the FAISS library for efficient similarity search. Experimental results demonstrate a clear performance-efficiency trade-off: ResNet101 achieved the highest classification accuracy, while EfficientNetB0 and MobileNetV2 offered significant advantages in inference speed. A user-friendly web interface was developed to support practical image retrieval applications. These findings highlight the potential of deep learning in enhancing the management and integrity of scientific image resources.

An attempt at fall detection on an embedded device based on YOLOv8n-pose

Dinh-Tu Nguyen, Loc-Dinh Tran, Van-Minh Huynh, Hoai-Tan Nguyen, Chi-Ngon Nguyen
Abstract | PDF
Human-action recognition aims to identify the actions performed by individuals. Due to the broad spectrum of human activities, action recognition covers a wide range. Among all, fall detection is a critical aspect of surveillance, particularly in environments where individuals are at risk. Throughout the years, several sensors, data types, and classification techniques have been investigated to address this issue. This paper proposes a lightweight fall-detection system that processes image sequences in real time. This system is deployed on an embedded device, specifically the Jetson Nano. Our goal is to construct a comprehensive dataset that accurately detects falls in various lighting conditions. The proposed system is constructed using YOLOv8n-pose, which has been trained to identify people using widely recognized dataset. Our methodology includes the design and implementation of YOLOv8n-pose, data collection, and rigorous testing to ensure the accuracy of real-time fall detection using a surveillance camera. The experimental results show that high detection over 90% accuracy and acceptable timing capabilities are achieved.

A review of evolving trends in construction project management: Integrating technology, leadership, and sustainability

Flory Mae Cari Cari, Edcel Comaingking, Alden Q. Gabuya Jr.
Abstract | PDF
This literature review synthesises recent advancements and emerging trends in the field of project management, with a focus on the evolving impact of technology, leadership, and sustainability. Drawing on a wide range of scholarly articles and industry publications, the review explores the foundational principles of project management, such as risk management, leadership effectiveness, and project efficiency, while also highlighting the complexities of applying these practices across various sectors. The integration of cutting-edge technologies, such as artificial intelligence, data analytics, and hybrid methodologies, is identified as a key driver of change in project execution. Furthermore, the increasing emphasis on sustainability and ethics reflects a shift in the industry toward a broader definition of project success, beyond traditional financial outcomes. This review emphasizes the need for continuous adaptation in project management practices and calls for future research to refine leadership strategies, improve technology adoption, and develop more holistic frameworks for assessing project success. The findings underscore the importance of aligning traditional project management principles with the demands of an increasingly complex and fast-evolving global environment.

Investigation of optical properties of Mn4+ doped ZnAl2O4 phosphor fabricated by sol-gel method

Ly Trieu Loc, Tran Thi Thuy Tien, Lam Tan Phat, Nguyen Tran Yen Vy, To Vo Hoai Phong, Nguyen Thi Kim Chi, Dinh Manh Tien
Abstract | PDF
In this work, ZnAl2O4: Mn4+ phosphor is prepared by the sol-gel method. The phase structure was investigated by X-ray diffraction (XRD), and morphology and element composition were analysed by scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDX), luminescent performance of the phosphor using photoluminescence (PL) spectra, and Raman spectroscopy. The obtained powder shows a red broad emission band peaking at 676 nm and 692 nm, corresponding to the energy transfer of electrons from 2Eg to 4A2g. The Mn4+ concentration is doped at 0.5 mol% into the ZnAl2O4 host lattice. After that, the powder annealed at 1200℃ for 4 hours in the air revealed an average crystallite size of about 52.833 nm and a high emission intensity with (x, y) CIE colour coordinates (0.2705, 0.2741). Therefore, ZnAl2O4: Mn4+ phosphor could be a promising material for solid-state lighting applications.

Vascular plant resources' contribution to the Vam Ho Bird Sanctuary and surrounding areas' designation as a Natural Heritage Site

Nguyen Thi Lan Thi, Nguyen Khanh Trinh Tram, Quang Dien Tran
Abstract | PDF
The survey of plant diversity and vegetation in the Vam Ho Bird Sanctuary and surrounding areas (Ba Tri and Binh Dai districts, Ben Tre province) was conducted in October 2024. The study aimed to build the plant diversity basics for investing and upgrading Vam Ho to the status of a Natural Heritage site. Using transect surveys and sampling, the study recorded 179 species belonging to 55 families of vascular plants. Among them, Azima sarmentosa, Oryza rufipogon, Psilotum nudum were listed in the Vietnam Red Book (2024), Avicennia marina var. rumphiana, Phoenix paludosa, Sonneratia ovata, Ceriops decandra were on the IUCN Red List (2025). In addition, a total of 158 out of 179 recorded species have had at least one of six main use groups: medicine, food, animal/poultry feed, ornamental plants, timber/charcoal/construction materials, and household items/handicrafts. The vegetation was divided into three types: the natural habitat in Vam Ho Bird Sanctuary, the Casuarina equisetifolia coastal plantation, and the estuarine-coastal mangrove forest. The establishment of the Natural Heritage Site will create positive conditions for the activities of habitat conservation and coastal protection, which help to reduce the pressure of environmental degradation on the Ba Lai estuary in the future.

Diversity of phytoplankton species in aquaculture ponds of Dak Ha, Sa Thay, Kon Plong districts, Quang Ngai province

Pham Thi Tam, Le Tri Khai, Doan Van Quang, Dang Thi Thuy, Le Thi Thom, Ngo Thi Hoai Thu, Nguyen Cam Ha, Le Anh Huy, Nguyen Manh Dat, Nguyen Thuy Lien, Dang Diem Hong
Abstract | PDF
This paper studied the diversity of phytoplankton species at 9 sampling points of aquaculture ponds in Dak Ha, Sa Thay and Kon Plong districts, Kon Tum province (now Quang Ngai province) in 2022–2023. The results recorded 7 phyla, 98 genera, and 142 species, of which 22 genera and 23 species of phytoplankton were present in all 3 sampling periods. The values of the biodiversity index (H’) and the diversity value index (Dv) at 9 sampling points ranged from 2.51–3.66 and 2.39–3.58, indicating that the diversity of phytoplankton in aquaculture ponds is high and very high, respectively. The regulation index (J) ranged from 0.95 to 0.99, indicating that phytoplankton species were evenly distributed at all studied sampling points. The similarity coefficient of species composition of phytoplankton populations at 9 sampling points ranged from 0.43 - 0.85. In water samples collected at 9 study locations, the number of phytoplankton species was negatively correlated with impedance. In particular, at the sampling locations, the main environmental factors such as suspended solids, salinity, and conductivity affecting the cell density of phytoplankton communities had a positive correlation and a negative correlation with impedance.

EMI Lecturers in Vietnamese Higher Education: Exploring professional development needs for teaching quality enhancement

Au Xuan Sam, Phuong Hoang Yen, Nguyen Thi Huong Lan
Abstract | PDF
As English-medium instruction (EMI) programs have expanded rapidly in Vietnam as a result of educational globalisation, more research is needed in this field. This study aims to promote EMI lecturers’ professional development and enhance teaching quality by identifying the needs of content lecturers in Vietnamese higher education, in order to gain a more comprehensive insight into appropriate support and policy development. Data were collected via questionnaires distributed to EMI lecturers (n=33) and in-depth interviews with six participants. The findings show that EMI lecturers required professional development support, particularly in improving their pedagogical competencies and English language proficiency. They also expressed a desire for collaboration with colleagues. In addition, this study identified lecturers' specific needs for institutional support, with particular emphasis on teaching materials and resources. These insights contribute to the existing literature and provide helpful suggestions for supporting EMI lecturers in similar settings.

Factors influencing online learning motivation of students at Can Tho University

Do Ngoc Thanh Thuy, Doan Thi Kieu My
Abstract | PDF
This study explores the factors influencing students' motivation for online learning at Can Tho University. The research examines five key factors: personal, lecturer-related, institutional, academic, and environmental aspects. A survey was conducted with 892 students across various academic disciplines. The study employed statistical analyses, including Cronbach’s Alpha reliability testing, exploratory factor analysis (EFA), and multiple regression analysis to determine the impact of these factors on students' motivation. The findings indicate that personal and lecturer-related factors have the most significant positive influence, highlighting the importance of self-discipline, time management, and lecturer support. Institutional factors also play a crucial role, particularly in terms of learning infrastructure and support services. However, environmental factors negatively affect motivation, such as poor internet connectivity, financial difficulties, and distractions, which hinder students’ engagement. Additionally, students with higher academic performance and greater online learning experience show stronger motivation. Differences across academic disciplines suggest the need for tailored teaching methods and institutional support. These findings provide insights into enhancing online learning motivation and contribute to policy recommendations improving the quality of online education at Can Tho University.

Gender and development concepts and principles in English 10 learning materials

Lea Ann A. Villanueva
Abstract | PDF
This study examined how gender and development (GAD) concepts and principles are incorporated into English 10 learning materials using a mixed-methods approach. The examination of lesson plans and self-learning modules through content analysis revealed biases supporting traditional gender roles, such as inadequate representation of women and showing stereotypical gender behaviours. Interviews with English 10 public teachers highlighted diverse strategies they employed, despite limited exposure to GAD concepts, suggesting a lack of comprehensive training in this area. However, teachers still endeavour to integrate these principles into their instruction to foster inclusivity, though challenges persist, such as biases influencing disciplinary measures. The findings of this study highlight the importance of challenging narrow views and promoting inclusive practices in educational materials. This study contributes to the ongoing efforts to promote gender equality and diversity in education highlighted in UNESCO’s Education for All initiative (2003), emphasizing the need for continuous awareness, sensitivity, and proactive integration of GAD principles in educational settings.

Education 4.0 technologies for English language teaching: A meta-synthesis

Reynarose Garado-Trasmil, Rivika C. Alda
Abstract | PDF
This meta-synthesis aimed to determine the technological landscape of English Language Teaching (ELT) in Asia over the past decade. Through meticulous screening of relevant studies, a thematic analysis of 14 qualitative and primary sources revealed a reliance on internet platforms, mobile devices, gadgets, and robots to facilitate effective English instruction. Yet, a critical evaluation of contemporaneity revealed deficiencies in ELT technologies, ranging from source unreliability to technological unfamiliarity, alongside notable strengths such as user-friendliness and timeliness of information. Addressing these shortcomings demands a concerted effort to equip students and educators with the requisite devices and training. Furthermore, fostering a conducive learning environment mandates tailored curriculum enhancements to bolster technical proficiency and the seamless integration of emerging technologies. Collaboration among parents, institutional leaders, and the educational community is paramount for navigating the adoption of new technologies, both within and beyond the traditional classroom. Delineating clear learning objectives, providing adequate skills training, and instituting appropriate curriculum changes are pivotal in harnessing the full potential of technology for English language learning.

A project evaluation on parent-teacher partnership for flexible learning

Charmane T. Espejo, Dhana A. Canoy, Pauline Giselle L. Paraan, Darryl Roy T. Montebon
Abstract | PDF
This paper aimed to evaluate a parent-teacher partnership project implemented in one laboratory school in Metro Manila. Parent-teacher partnership is a collective effort by parents to work together for the benefit of learners. A project aims to address a particular problem and can be an intervention which ends once it has been implemented and regularly used. Participants are middle-income parents. Convenient sampling was used to identify the participants. After planning and designing a project in support of parent-teacher partnership, it was implemented to equip the parents with the necessary strategies to assist their children in flexible learning. After close collaboration with the parents, the researchers evaluated the project using a survey questionnaire. The results show that the objectives of the project were met, and the content was helpful in supporting parent-teacher partnerships. Also, relevant findings revealed that parents perceived the project as informative, as they gained strategies and tools to support their child's academic learning and were equipped with socio-emotional strategies to manage their child’s socio-emotional needs and to teach them how to navigate the educational technologies used in the school. The results of the study provided insights into revisiting policies to engage parents so they can develop...

Evaluation of antimicrobial activity from the Tetragonula sp. propolis extract

Tran Thanh Toi, Huynh Vy, Do Tan Khang, Nguyen Pham Anh Thi, Tran Ngoc Quy
Abstract | PDF
The aim of this study was to investigate the antimicrobial activity against Staphylococcus aureus, Cutibacterium acnes, Escherichia coli, and Candida albicans of the ethanol extract and the n-hexane and ethyl acetate extract of the stingless bee Tetragonula sp. propolis. Additionally, the study also identified the stingless bee species based on morphological characteristics and Cytochrome c oxidase subunit I (COI) gene sequence. The antimicrobial activity of three stingless bee propolis extracts showed resistance against all three bacterial S. aureus, C. acnes, and E. coli. At a concentration of 1000 μg/mL, the ethanol extract exhibited the highest resistance, with C. acnes showing a zone of inhibition of 13.33 mm, while the inhibition zones were 11.33 mm for S. aureus and 10.33 mm for E. coli. The Minimum Inhibitory Concentration (MIC) for the three bacterial S. aureus, C. acnes, and E. coli were found to be 30, 50, and 50 μg/mL, respectively. The study identified the presence of polyphenols, flavonoids, aglycones, phenolics, and ketones quantified the total phenolic and flavonoid content, assessed the antibacterial activity of the propolis extract against the three pathogenic bacteria, and provided the foundation for further research.

A robust ensemble framework for helmet usage classification in real-world scenarios

Tan-Duy Lam, Tuong Le
Abstract | PDF
The application of machine learning models in the analysis of helmet-related images has yielded remarkable results in identifying and classifying helmet-wearing behaviours. Previous research has employed several pretrained models to predict proper or improper helmet use, achieving high accuracy on the Helmet Wearing Image Dataset (2024), a newly introduced dataset designed to enhance classification capabilities. This study aims to improve prediction performance on helmet datasets by leveraging state-of-the-art deep learning models and ensemble techniques. Using ResNet-50, MobileNetV2, and EfficientNet-B0 models, the proposed EnsemHelmet Framework uses a soft voting ensemble to optimise the classification results, achieving an outstanding accuracy of 99.24% on the experimental dataset. The results demonstrate the potential of ensemble learning to achieve high performance. This study not only improves the accuracy of the helmet-wearing recognition system but also highlights the effectiveness of ensemble techniques in optimizing performance on real-world datasets.
View All Issues