IIUM Engineering Journal
Not a member yet
783 research outputs found
Sort by
Proposed ConvBiLSTM-Net Model for Enhancing Earthquake Prediction Performance Using Spatiotemporal Features
Full text linkAccurate earthquake prediction remains a significant challenge due to the complex spatiotemporal dependencies inherent in seismic events. To address this issue, the present study proposes ConvBiLSTM-Net. This hybrid deep learning model combines Convolutional Neural Networks (CNNs) for spatial feature extraction with Bidirectional Long Short-Term Memory (BiLSTM) networks for temporal sequence modeling. The model integrates historical earthquake data with spatial information in the form of fault density (FD), derived using Kernel Density Estimation (KDE). The KDE bandwidth is optimized using the Bivariate Local Indicator of Spatial Association (LISA) method to enhance spatial adaptivity. The dataset comprises earthquake records from the USGS catalog (1974–2023) and active fault data compiled in the 2017 Indonesian Earthquake Source and Hazard Map, published by the National Earthquake Study Center (PuSGeN). ConvBiLSTM-Net is evaluated under short-term and medium-term prediction scenarios, targeting earthquake magnitude, depth, and epicenter coordinates (latitude and longitude), using standard performance metrics such as accuracy, F1 score, root mean square error (RMSE), mean absolute error (MAE), and the coefficient of determination (R²). In the short-term scenario, the model achieves average improvements of 9.31% in R², 3.41% in accuracy, and 6.06% in F1 score, while reducing RMSE by 10.63% and MAE by 12.40% across magnitude, depth, and latitude predictions. For longitude, R², accuracy, and F1 score also improve by 10.88%, 11.76%, and 17.54%, respectively, although RMSE and MAE increase by 13.09% and 20.74%, indicating a trade-off between enhanced pattern recognition and higher absolute error. Under the medium-term scenario, the model demonstrates average improvements of 7.49% in R², 3.39% in accuracy, and 7.06% in F1 score, while reducing RMSE and MAE by 6.22% and 17.72%, respectively, for magnitude, depth, and latitude predictions. For longitude, R², accuracy, and F1 score improve by 12.50%, 2.48%, and 1.60%, respectively, though RMSE and MAE increase by 37.31% and 37.01%, again highlighting a trade-off between better pattern recognition and increased absolute error in this dimension. These findings demonstrate that ConvBiLSTM-Net, engineered to integrate spatial and temporal features, is a robust and adaptive architecture for enhancing earthquake prediction performance. Its spatiotemporal modeling approach yields consistently high accuracy and stability across forecasting horizons, particularly in predicting earthquake epicenters. Despite minor trade-offs in absolute error for longitude predictions, the overall performance improvements affirm its potential as a reliable tool for seismic hazard assessment and disaster risk mitigation.
ABSTRAK: Ramalan gempa bumi yang tepat kekal sebagai satu cabaran utama disebabkan oleh kebergantungan spatiotemporal yang kompleks dalam kejadian seismik. Bagi menangani isu ini, kajian ini mencadangkan ConvBiLSTM-Net, iaitu sebuah model hibrid pembelajaran mendalam yang menggabungkan Rangkaian Neural Konvolusional (CNN) dan Memori Jangka Pendek Dwi Arah (BiLSTM), bagi tujuan pengekstrakan ciri spatial dan pemodelan jujukan temporal, masing-masing. Model ini menggabungkan data sejarah gempa bumi dengan maklumat spatial dalam bentuk ketumpatan sesar (fault density, FD), yang diperoleh melalui Kaedah Anggaran Ketumpatan Kernel (Kernel Density Estimation, KDE). Lebar jalur KDE dioptimumkan menggunakan kaedah Bivariate Local Indicator of Spatial Association (LISA) bagi meningkatkan kepekaan spatial. Set data kajian merangkumi rekod gempa bumi daripada katalog USGS (1974–2023) serta data sesar aktif yang disusun dalam Peta Sumber dan Bahaya Gempa Indonesia 2017, terbitan Pusat Kajian Gempa Nasional (PuSGeN). Model ConvBiLSTM-Net ini dinilai dalam dua senario ramalan—jangka pendek dan jangka sederhana—bagi parameter magnitud, kedalaman, serta koordinat pusat gempa (latitud dan longitud), dengan menggunakan metrik standard seperti ketepatan, skor F1, RMSE, MAE dan pekali penentuan (R²). Malalui senario jangka pendek, model mencatatkan purata peningkatan sebanyak 9.31% pada R², 3.41% pada ketepatan, dan 6.06% pada skor F1; serta pengurangan RMSE sebanyak 10.63% dan MAE sebanyak 12.40% merentas ramalan magnitud, kedalaman, dan latitud. Bagi dimensi longitud, R², ketepatan, dan skor F1 turut meningkat sebanyak 10.88%, 11.76%, dan 17.54% masing-masing; namun begitu, RMSE dan MAE meningkat sebanyak 13.09% dan 20.74%, menunjukkan kompromi antara pengecaman corak yang lebih baik dan ralat mutlak yang lebih tinggi. Manakala senario jangka sederhana, model mencatatkan purata peningkatan sebanyak 7.49% pada R², 3.39% pada ketepatan, dan 7.06% pada skor F1; serta pengurangan RMSE sebanyak 6.22% dan MAE sebanyak 17.72% merentas tugas ramalan magnitud, kedalaman, dan latitud. Bagi dimensi longitud, peningkatan masing-masing dicatatkan pada R² (12.50%), ketepatan (2.48%), dan skor F1 (1.60%), tetapi RMSE dan MAE meningkat secara ketara sebanyak 37.31% dan 37.01%, menunjukkan kompromi antara pengecaman corak yang lebih kukuh dan ralat mutlak yang lebih besar pada dimensi ini. Dapatan kajian ini membuktikan bahawa ConvBiLSTM?Net, yang direka bentuk bagi menggabungkan ciri-ciri spatial dan temporal, merupakan satu seni bina model yang teguh dan adaptif dalam meningkatkan prestasi ramalan gempa bumi. Pemodelan spatiotemporal bersepadu yang digunakan menghasilkan tahap ketepatan dan kestabilan yang tinggi secara konsisten merentasi pelbagai ufuk ramalan, terutamanya dalam menentukan lokasi pusat gempa bumi. Walaupun terdapat sedikit kekurangan dalam nilai ralat mutlak bagi ramalan longitud, peningkatan prestasi secara keseluruhan mengesahkan nilainya sebagai alat yang boleh dipercayai dalam penilaian bahaya seismik dan pengurangan risiko bencana
Editorial
Full text linkThe IIUM Engineering Journal Vol. 26 No. 3 continues its commitment to advancing impactful research that addresses contemporary engineering challenges through innovation, sustainability, and interdisciplinary integration. This issue features 28 high-quality papers across a broad spectrum of disciplines: Civil and Environmental Engineering, Electrical, Computer and Communications Engineering, Engineering Mathematics and Applied Science, Materials and Manufacturing Engineering, and Mechatronics and Automation Engineering.
The Civil and Environmental Engineering section reflects a growing focus on safety, durability, and sustainable infrastructure. Papers in this category explore accident probability among elderly motorcyclists in Indonesia, corrosion assessment in fly ash–silica fume concrete via non-destructive testing, eco-processed pozzolan characterization, strength forecasting using machine learning and response surface methodology, clayey sand improvement using brick dust, and sediment scour simulations in submerged weirs. Collectively, these works contribute to improved construction practices and enhanced environmental resilience.
The Electrical, Computer, and Communications Engineering section comprises the largest set of contributions, showcasing cutting-edge AI and computational advancements across multiple domains. Topics include real-time deep learning for UAV object detection, electronic nose integration with LeNet and regularization, blockchain benchmarking on heterogeneous IoT hardware, and health monitoring of lead-acid batteries in off-grid solar systems. Other notable studies investigate fake news detection using multimodal embeddings, hybrid deep learning for data center temperature control, sensor-based fingerspelling recognition, and spatiotemporal models for earthquake prediction. Additionally, papers on chipless RFID resonators, LoRa-driven UAV return prediction, saliency-based segmentation of skin cancer, and sentiment analysis on TVET from social media highlight AI’s broad impact. Further innovations are presented in malware detection within IoMT environments, trade-space analysis of satellite anomalies, parametric cost modeling, and power generation optimization using bio-inspired algorithms. These papers demonstrate the transformative role of AI, IoT, and intelligent systems in reshaping engineering practice.
In Engineering Mathematics and Applied Science, a significant theoretical contribution explores a novel seven-dimensional chaotic system, enriching our understanding of nonlinear dynamics and complex systems. The Materials and Manufacturing Engineering section highlights progress in safety engineering and the development of advanced composites. Featured studies include the enhanced design of portable oil spill skimmers using computational fluid dynamics, void formation analysis in BFS/CaCO? diffusion couples, and the evaluation of machining performance in hybrid fiber-reinforced polymer composites. These contributions reflect ongoing efforts to optimize material performance and environmental protection.
In Mechatronics and Automation Engineering, contributions emphasize intelligent control and autonomous systems. Topics include the dynamic analysis of aerial work platforms using hybrid CAD for satellite testing, as well as autonomous navigation based on frontier-based detection integrated with social force modeling. These efforts highlight the integration of robotics, control systems, and intelligent automation to improve operational reliability and mobility in complex environments.
This issue exemplifies the journal’s mission to bridge theory and real-world application across diverse engineering domains, promoting secure, sustainable, and intelligent systems. The editorial team expresses its sincere gratitude to all authors, reviewers, and section editors for their valuable contributions, commitment to excellence, and support in upholding the scholarly integrity of this publication. We hope these contributions inspire future innovations, foster stronger academia–industry collaboration, and accelerate global progress towards sustainable, inclusive, and technology-driven development.
Prof. Ir. Ts. Dr. Teddy Surya GunawanExecutive EditorIIUM Engineering Journal
ISSN: 1511-788X E-ISSN: 2289-7860
Published by:IIUM Press,International Islamic University MalaysiaJalan Gombak, 53100 Kuala Lumpur, MalaysiaPhone (+603) 6421-5014, Fax: (+603) 6421-629
Global-Local Self-Attention-Based Long Short-Term Memory with Optimization Algorithm for Speaker Identification
Full text linkSpeaker identification (SI) involves recognizing a speaker from a group of unknown speakers, while speaker verification (SV) determines if a given voice sample belongs to a particular person. The main drawbacks of SI are session variability, noise in the background, and insufficient information. To mitigate the limitations mentioned above, this research proposes Global Local Self-Attention (GLSA) based Long Short-Term Memory (LSTM) with Exponential Neighborhood – Grey Wolf Optimization (EN-GWO) method for effective speaker identification using TIMIT and VoxCeleb 1 datasets. The GLSA is incorporated in LSTM, which focuses on the required data, and the hyperparameters are tuned using the EN-GWO, which enhances speaker identification performance. The GLSA-LSTM with EN-GWO method acquires an accuracy of 99.36% on the TIMIT dataset, and an accuracy of 93.45% on the VoxCeleb 1 datasets, while compared to SincNet and Generative Adversarial Network (SincGAN) and Hybrid Neural Network – Support Vector Machine (NN-SVM).
ABSTRAK: Pengenalpastian pembicara (Speaker Identification, SI) melibatkan pengenalan pembicara daripada kumpulan pembicara yang tidak dikenali, manakala pengesahan pembicara (Speaker Verification, SV) menentukan sama ada sampel suara tertentu milik seseorang individu. Kekurangan utama dalam SI ialah variasi sesi, bunyi latar belakang, dan maklumat yang tidak mencukupi. Untuk mengatasi kekangan tersebut, kajian ini mencadangkan kaedah Global Local Self-Attention (GLSA) berasaskan Long Short-Term Memory (LSTM) dengan Pengoptimuman Grey Wolf Jiranan Eksponen (EN-GWO) bagi pengenalpastian pembicara yang berkesan menggunakan set data TIMIT dan VoxCeleb 1. GLSA digabungkan dalam LSTM yang memberi tumpuan pada data yang diperlukan, manakala parameter hiper ditala menggunakan EN-GWO untuk meningkatkan prestasi pengenalpastian pembicara. Kaedah GLSA-LSTM dengan EN-GWO mencapai ketepatan 99.36% pada dataset TIMIT dan ketepatan 93.45% pada dataset VoxCeleb 1, berbanding dengan SincNet dan Generative Adversarial Network (SincGAN) serta Hybrid Neural Network – Support Vector Machine (NN-SVM)
Ozonation of Vegetable Oils and Study on Their Physicochemical and Biological Characteristics
Full text linkFree ozone offers significant benefits in biological applications due to its efficacy as a disinfectant, but toxicity and instability are associated with it. Hence, producing ozonated vegetable oil (OVO) has been explored as a potential solution, yielding stable ozonation by-products with medical potential, such as antimicrobial activity. Several studies have explored OVO's characteristics and biological effects, including olive oil, sunflower oil, and canola oil. However, optimizing ozonation conditions is still lacking, with many other types of vegetable oils yet to be studied. This research comprises three phases: i) ozonation of selected oils: red palm oil (RPO), rice bran oil (RBO), peanut oil (PO), and virgin coconut oil (VCO), ii) screening for the most effective OVO against three bacteria (Staphylococcus aureus, Bacillus subtilis, and Escherichia coli), and iii) physicochemical testing. Results show increased peroxide and acidity values in most OVO and a decrease in iodine value compared to untreated oil. Ozonated virgin coconut oil (OVCO) exhibits the highest antibacterial activity by showing a zone of inhibition of 11.3 mm and 84.35% killing rate at 30 minutes incubation time, particularly against S. aureus. Further optimization using Design Expert®6.0.8 software identifies the most effective ozonation conditions for OVCO, achieving a peak killing rate of 100% against S. aureus with 360 mins of ozone exposure and ozone flow rates of 1 l/min. Kinetic studies confirm rapid bacterial eradication, with over 90% of S. aureus killed by OVCO within 2 mins. Moreover, OVCO proved to be non-toxic to human foreskin fibroblast (HFF1) cells, maintaining 80% viability even after exposure to 1 mg/ml OVCO treated with ozone for 120 and 240 mins. These findings underscore the promising medical potential of OVCO, particularly in treating skin diseases.
ABSTRAK: Ozon bebas menawarkan manfaat signifikan dalam aplikasi biologi disebabkan keberkesanannya sebagai bahan pembasmi kuman, namun ia turut dikaitkan dengan ketoksikan dan ketidakstabilan. Oleh itu, penghasilan minyak sayuran berozon (OVO) telah diteroka sebagai potensi penyelesaian, menghasilkan hasil sampingan ozonasi yang stabil dengan potensi perubatan seperti aktiviti antimikrob. Beberapa kajian telah meneliti ciri-ciri dan kesan biologi OVO termasuk minyak zaitun, minyak bunga matahari, dan minyak kanola. Namun, proses pengoptimuman keadaan ozonasi masih belum lengkap dan banyak lagi jenis minyak sayuran belum dikaji. Kajian ini terdiri daripada tiga fasa: i) ozonasi minyak terpilih iaitu minyak sawit merah (RPO), minyak dedak padi (RBO), minyak kacang tanah (PO), dan minyak kelapa dara (VCO), ii) saringan keberkesanan OVO terhadap tiga jenis bakteria (Staphilokokus aureus, Basillus subtilis, dan Escherichia coli), dan iii) ujian fisikokimia. Keputusan menunjukkan peningkatan nilai peroksida dan keasidan dalam kebanyakan OVO serta penurunan nilai iodin berbanding minyak yang tidak dirawat. Minyak kelapa dara berozon (OVCO) menunjukkan aktiviti antibakteria tertinggi dengan zon perencatan berdiameter 11.3 mm dan kadar pembunuhan bakteria sebanyak 84.35% dalam masa inkubasi 30 minit, khususnya terhadap S. aureus. Pengoptimuman lanjut menggunakan perisian Design Expert®6.0.8 mengenal pasti keadaan ozonasi paling berkesan bagi OVCO, dengan pencapaian kadar pembunuhan maksimum 100% terhadap S. aureus pada pendedahan ozon selama 360 minit dan kadar aliran ozon 1 l/min. Kajian kinetik mengesahkan penghapusan bakteria yang pantas, dengan lebih 90% S. aureus dibunuh oleh OVCO dalam masa 2 minit. Tambahan, OVCO terbukti tidak toksik terhadap sel fibroblas kulit manusia (HFF1), dengan mengekalkan 80% daya hidup walaupun selepas pendedahan kepada 1 mg/ml OVCO yang dirawat ozon selama 120 dan 240 minit. Penemuan ini menekankan potensi perubatan OVCO, khususnya dalam merawat penyakit kulit
Boundary-Aware Saliency-Based Level Set with Momentum Contrast Metaformer for Skin Cancer Segmentation and Classification
Full text linkSkin cancer is considered one of the most widespread and life-threatening cancers and remains a challenging task for dermatologists. This challenge arises due to the small boundaries and regions of affected tissue. Existing methods yield poor classification accuracy due to inefficient classifier performance in recognizing complex patterns. Therefore, an effective skin cancer segmentation and classification method called Boundary-Aware Saliency-based Level Set (BASLS) with Momentum Contrast Metaformer (MC-Metaformer) is proposed in this research. BASLS enables improved lesion segmentation by identifying significant structural components along lesion edges. Using residual connections, MC-Metaformer provides a better gradient path, addressing gradient fading issues during deep feature extraction. Preprocessing is performed on the HAM10000, ISIC-2019, and ISIC-2020 datasets to improve and standardize the data. ResNet50 is used to extract relevant features for classification. Experimental results demonstrate that the proposed MC-Metaformer outperforms the Deep Convolutional Neural Network (DCNN), achieving classification accuracies of 99.58%, 99.32%, and 98.62% on the HAM10000, ISIC-2019, and ISIC-2020 datasets, respectively. These results confirm the robustness and efficiency of the model in accurate skin cancer segmentation and classification.
ABSTRAK: Kanser kulit merupakan salah satu jenis kanser paling meluas dan mengancam nyawa serta masih menjadi cabaran utama kepada pakar dermatologi. Cabaran ini berpunca daripada sempadan dan kawasan kecil tisu yang terjejas. Kaedah sedia ada mempunyai ketepatan klasifikasi yang rendah kerana prestasi pengelasan yang kurang berkesan dalam mengenal pasti corak kompleks. Oleh itu, kajian ini mencadangkan kaedah segmentasi dan klasifikasi kanser kulit yang berkesan dikenali sebagai Boundary-Aware Saliency-based Level Set (BASLS) bersama Momentum Contrast Metaformer (MC-Metaformer). BASLS membolehkan segmentasi lesi yang lebih baik dengan mengenal pasti komponen struktur penting di sepanjang tepi lesi. Melalui penggunaan sambungan residu, MC-Metaformer menyediakan laluan kecerunan yang lebih baik bagi menangani isu kehilangan kecerunan semasa pengekstrakan ciri mendalam. Pra-pemprosesan dilakukan ke atas set data HAM10000, ISIC-2019, dan ISIC-2020 bagi meningkatkan serta menyeragam data. ResNet50 digunakan bagi mengekstrak ciri relevan untuk klasifikasi. Dapatan eksperimen menunjukkan bahawa MC-Metaformer yang dicadangkan mengatasi Rangkaian Neural Konvolusi Mendalam (DCNN) dengan ketepatan klasifikasi sebanyak 99.58%, 99.32%, dan 98.62% masing-masing pada set data HAM10000, ISIC-2019, dan ISIC-2020. Dapatan ini mengesahkan keteguhan dan kecekapan model dalam segmentasi dan klasifikasi kanser kulit secara tepat
Machine Learning and RSM for Strength Forecasting in Sustainable SCGC
Full text linkThis research focuses on the predictive modeling of flexural (Ff) and splitting tensile (Ft) strengths in Self-Compacting Geopolymer Concrete (SCGC) to support sustainable mix design optimization. A curated dataset comprising 544 experimental records was utilized to train and evaluate eight supervised machine learning (ML) algorithms. These included Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Random Forests, Gradient Boosting, CN2 Rule Induction, Naïve Bayes, Decision Trees, and Stochastic Gradient Descent. The predictive performance of each model was assessed using multiple statistical metrics, such as RMSE, R², and accuracy percentage. Among the models, SVM and KNN achieved the highest precision, with R² values of 0.99 and RMSE as low as 0.10 MPa. Additionally, statistical techniques were applied to identify influential input variables, confirming the dominant role of binder constituents in determining tensile-related strength. The models demonstrated strong generalization on unseen data and minimal sensitivity to activator dosage or curing age. These results validate the effectiveness of ML-driven tools for SCGC prediction and offer a scalable framework for integrating data analytics into sustainable concrete design and performance optimization.
ABSTRAK: Kajian ini memfokuskan kepada pemodelan ramalan bagi kekuatan lenturan (Ff) dan tegangan belahan (Ft) dalam Konkrit Geopolimer Pemadat Kendiri (SCGC) bagi menyokong pengoptimuman reka bentuk campuran mampan. Satu set data terpilih yang merangkumi 544 rekod eksperimen telah digunakan bagi melatih dan menilai lapan algoritma pembelajaran mesin (ML) terselia. Algoritma tersebut termasuk Mesin Sokongan Vektor (SVM), K-Nearest Neighbors (KNN), Rawak Forests, Gradient Boosting, CN2 Rule Induction, Naïve Bayes, Pokok Keputusan, dan Stochastic Gradient Descent. Prestasi ramalan setiap model dinilai menggunakan pelbagai metrik statistik seperti RMSE, R², dan peratusan ketepatan. Antara model tersebut, SVM dan KNN mencapai ketepatan tertinggi dengan nilai R² sebanyak 0.99 dan RMSE serendah 0.10 MPa. Tambahan, teknik statistik turut digunakan bagi mengenal pasti pemboleh ubah input berpengaruh, sekali gus mengesahkan peranan dominan konstituen pengikat dalam menentukan kekuatan berkaitan tegangan. Model yang dibangunkan menunjukkan keupayaan generalisasi yang kukuh terhadap data baharu serta kepekaan minimum terhadap dos pengaktif atau umur pengerasan. Dapatan ini mengesahkan keberkesanan alat berasaskan ML bagi meramal SCGC dan menawarkan kerangka boleh skala bagi mengintegrasikan analitik data ke dalam reka bentuk konkrit mampan serta pengoptimuman prestasi
Physical and Mechanical Properties of Green Cementless Mortar Incorporating Waste Paper Sludge Ash
Full text linkAs an alternative to conventional construction material such as normal concrete, waste paper sludge ash (WPSA) based geopolymer is seen as a promising and viable option in construction material selection due to its high amount of aluminum (Al) and silicon (Si) content. This research aims to determine the microstructure and mechanical characteristics of WPSA in geopolymers. The alkaline solution that contains 6 M of sodium hydroxide (NaOH) and sodium silicate (Na2SiO3) was used to activate the geopolymer. The hardened 50 mm-sized mortars were prepared and underwent a heat-cured process for 1 day at various temperatures at 24 ?, 60 ?, and 90 ?, respectively. Then, the mortar cubes were placed in the laboratory until the testing days. A compression test was conducted to identify the strength development of the WPSA-based geopolymer mortar at 7, 14, and 28 days, respectively. Chemical composition was analyzed using X-ray fluorescence (XRF). Furthermore, Fourier-transformed infrared spectroscopy (FTIR) was conducted to ascertain the structural elucidation and scanning electron microscope (SEM) analysis was done to provide microstructural observations of the geopolymer. Based on the XRF analysis, the WPSA has the highest amount of calcium oxide (CaO) instead of aluminum oxide (Al2O3) and silicon dioxide (SiO2), and it reduces the performance of WPSA as a cement replacement material. The ratio of SiO2 and Al2O3 is recorded as 1.1:1. Therefore, it is suitable for bricks and ceramics production instead of concrete production. As for the curing process, the heat-cured method is evident in accelerating strength development in the WPSA-based geopolymer mortar compared to the ambient curing method due to the rapid polymerization process in the geopolymer system. It is proven that 60 ? is the optimum temperature for the curing process for geopolymer mortar.
ABSTRAK: Sebagai salah satu alternatif kepada bahan binaan konvensional seperti konkrit biasa, geopolimer berasaskan abu enap cemar kertas (WPSA) adalah dilihat sebagai pilihan yang baik kerana bahan ini mempunyai kandungan aluminium (Al) and silika (Si) yang tinggi. Penyelidikan ini bertujuan untuk menentukan struktur mikro dan ciri mekanikal WPSA dalam geopolimer. Larutan alkali yang mengandungi 6 M natrium hidroksida (NaOH) dan natrium silikat (Na2SiO3) digunakan untuk mengaktifkan geopolimer. Mortar yang bersaiz 50 mm telah disediakn dan ia melalui proses pengawetan haba selama 1 hari pada suhu yang berza-beza, iaitu 24 ?, 60 ?, and 90 ?. Kemudian, kiub mortar tersebut diletakkan di dalam makmal sehingga hari ujian. Ujian mampatan dijalankan untuk mengenal pasti perkembangan kekuatan mortar geopolimer apabila mencapai 7, 14, dan 28 hari. Komposisi kimia dalam sampel telah dianalisa menggunakan pendarfluor sinar-X (XRF). Tambahan pula, Spektroskopi inframerah fourier transformasi (FTIR) untuk memastikan sifat bahan dan analisis pengimbasan mikroskop elektron (SEM) untuk menyediakan pemerhatian struktur mikro geopolimer. Berdasarkan data dari analisa XRF, WPSA mempunyai jumlah kalsium oksida (CaO) tertinggi berbanding aluminium oksida (Al2O3) dan silikon dioksida (SiO2), dan ia mengurangkan prestasi WPSA sebagai bahan gantian simen. Nisbah SiO2 dan Al2O3 direkodkan sebagai 1.1:1. Oleh itu, ia sesuai digunakan untuk pengeluaran batu bata dan seramik berbanding pengeluaran konkrit. Bagi proses pengawetan, kaedah pengawetan haba terbukti mempercepat perkembangan kekuatan geopolimer berasaskan WPSA berbanding pengawetan pada suhu ambien disebabkan berlakunya proses pempolimeran yang sangat pantas dalam sistem geopolimer tersebut. Telah terbukti bahawa suhu 60 ? adalah suhu optimum bagi proses pengawetan mortar geopolimer berasaskan WPSA ini
Map Floodwater Radar Imagery using Machine Learning Algorithms
Full text linkFlooding is a widespread and costly natural disaster around the world. Accurately assessing the extent of flooding in near real-time is crucial for governments and humanitarian organizations. This information strengthens early warning systems, evaluates risks, and guides effective relief efforts. Therefore, precise flood mapping is essential for saving lives through improved early warning systems and targeted emergency responses. In this study, radar imagery available on the Planetary Computer Data was utilized to train a U-Net model specifically designed to label flood-affected pixels in an image from a flood event. Different blocks of the U-Net encoder architecture were fine-tuned to identify the most efficient fine-tuned model, and their results were compared. As a result, the model with blocks 1 and 2 being fine-tuned demonstrated the highest Intersection over Union (IoU) score of 78.904%, an increase of 8.663% over the baseline methods.
ABSTRAK: Banjir merupakan bencana alam yang meluas dan mahal di seluruh dunia. Penilaian yang tepat terhadap skala banjir secara hampir masa nyata adalah penting bagi kerajaan dan organisasi kemanusiaan. Maklumat ini memperkukuhkan sistem amaran awal, menilai risiko, dan membimbing usaha bantuan yang lebih berkesan. Oleh itu, pemetaan banjir yang tepat adalah penting untuk menyelamatkan nyawa melalui sistem amaran awal yang lebih baik dan respons kecemasan yang disasarkan. Dalam kajian ini, imej radar yang tersedia pada Planetary Computer Data digunakan untuk melatih model U-Net yang direka khas untuk melabelkan piksel yang terjejas oleh banjir dalam imej daripada kejadian banjir. Bagi mengenal pasti model ditala-halus yang paling cekap, blok-blok berlainan dalam arkitektur pengekod U-Net telah ditala-halus, dan hasilnya dibandingkan. Hasilnya, model dengan blok 1 dan 2 yang ditala-halus menunjukkan skor Intersection over Union (IoU) tertinggi sebanyak 78.904%, iaitu peningkatan sebanyak 8.663% berbanding kaedah asas
Enhanced Early Autism Screening: Assessing Domain Adaptation with Distributed Facial Image Datasets and Deep Federated Learning
Full text linkThis study offers a significant advancement in the area of early autism screening by offering diverse domain facial image datasets specifically designed for the detection of Autism Spectrum Disorder (ASD). It stands out as the pioneering effort to analyze two facial image datasets – Kaggle and YTUIA, using federated learning methods to adapt domain differences successfully. The federated learning scheme effectively addresses the integrity issue of sensitive medical information and guarantees a wide range of feature learning, leading to improved assessment performance across diverse datasets. By employing Xception as the backbone for federated learning, a remarkable accuracy rate of almost 90% is attained across all test sets, representing a significant enhancement of more than 30% for the different domain test sets. This work is a significant and remarkable contribution to early autism screening research due to its unique novel dataset, analytical methods, and focus on data confidentiality. This resource offers a comprehensive understanding of the challenges and opportunities in the field of ASD diagnosis, catering to both professionals and aspiring scholars.
ABSTRAK: Kajian ini menawarkan kemajuan yang ketara dalam bidang saringan awal autisme dengan menyediakan pelbagai set data imej wajah yang direka khusus untuk pengesanan Gangguan Spektrum Autisme (ASD). Kajian ini menonjol sebagai usaha perintis untuk menganalisis dua set data imej wajah – Kaggle dan YTUIA, menggunakan kaedah pembelajaran teragih untuk menyesuaikan perbezaan domain dengan jayanya. Skim pembelajaran teragih ini berkesan menangani isu integriti maklumat perubatan sensitif dan menjamin pembelajaran ciri yang meluas, yang membawa kepada prestasi penilaian yang lebih baik merentas set data yang berbeza. Dengan menggunakan Xception sebagai tunjang pembelajaran teragih, kadar ketepatan yang luar biasa hampir 90% dicapai merentas semua set ujian, mewakili peningkatan ketara lebih daripada 30% untuk set ujian domain yang berbeza. Hasil kerja ini merupakan sumbangan penting dan luar biasa dalam penyelidikan saringan awal autisme kerana set data yang unik dan baharu, kaedah analisis yang digunakan, serta tumpuan kepada kerahsiaan data. Sumber ini menawarkan pemahaman yang menyeluruh mengenai cabaran dan peluang dalam bidang diagnosis ASD, sesuai untuk para profesional dan sarjana yang berminat
Development of 3D-Printed Serpentine Fluidic Channel Integrated with Heating Element for Loop-Mediated Isothermal Amplification (LAMP) Process
Full text linkDNA-based point-of-care (POC) diagnostics require rapid, accurate, and portable platforms for detection of infectious diseases. This can be achieved by incorporating a loop-mediated isothermal amplification (LAMP) process for DNA amplification into the system. LAMP offers a promising in-situ solution, but maintaining consistent reaction conditions, such as a constant temperature, specifically at 65°C for 35 minutes to complete the LAMP process, remains a critical challenge. Therefore, this work presents the development of a 3D-printed serpentine fluidic channel integrated with a heating element for DNA amplification through the LAMP process. To assess their heating capabilities, heating testing was initially performed on several commercially available heating elements (Heater Cartridge, PTC 140, and PTC 230). PTC 230 heating element was chosen for its rapid heating performance (reaching 65°C in 54.78 seconds). Later, three serpentine fluidic channels of different diameters (1.6 mm, 1.7 mm, and 1.8 mm) were fabricated using a Masked Stereolithography Apparatus (MSLA) 3D printer. The developed portable LAMP device consisting of a fabricated serpentine fluidic channel on a PTC 230 heating element allows the sample to be heated at 65°C for 35 minutes. Sample flow inside each serpentine fluidic channel was measured and compared with the expected flow time of 35 minutes. It was observed that the fluidic channel with a 1.6 mm diameter shows the closest value of 34.33 minutes (percentage deviation of 1.91%) as compared to the other two channels. The optimized fluidic channel design (channel diameter of 1.6 mm) coupled with the rapid heating performance of the PTC 230 element (reaching 65°C in 54.78 seconds) for a portable LAMP device represents a significant step towards developing rapid, accurate, and portable POC diagnostic tools.
ABSTRAK: Diagnostik point-of-care (POC) berasaskan DNA memerlukan platform yang pantas, tepat, dan mudah alih untuk mengesan penyakit berjangkit. Ini boleh dicapai dengan menggabungkan proses penguatan isoterma bersandar gelung (LAMP) ke dalam sistem untuk penguatan DNA. LAMP menawarkan penyelesaian in-situ yang menjanjikan, tetapi mengekalkan keadaan reaksi yang konsisten, seperti suhu tetap pada 65°C selama 35 minit untuk menyelesaikan proses LAMP, kekal sebagai cabaran kritikal. Oleh itu, kajian ini membentangkan pembangunan saluran bendalir berlingkar 3D yang dicetak dengan integrasi elemen pemanas untuk penguatan DNA melalui proses LAMP. Untuk menilai keupayaan pemanasannya, ujian pemanasan dijalankan pada beberapa elemen pemanas komersial yang tersedia (Heater Cartridge, PTC 140, dan PTC 230). Elemen pemanas PTC 230 dipilih kerana prestasi pemanasannya yang pantas (mencapai 65°C dalam 54.78 saat). Selepas itu, tiga saluran bendalir berlingkar dengan diameter berbeza (1.6 mm, 1.7 mm, dan 1.8 mm) telah dihasilkan menggunakan pencetak 3D Masked Stereolithography Apparatus (MSLA). Peranti LAMP mudah alih yang dibangunkan, terdiri daripada saluran bendalir berlingkar yang dihasilkan di atas elemen pemanas PTC 230, membolehkan sampel dipanaskan pada suhu 65°C selama 35 minit. Aliran sampel di dalam setiap saluran bendalir berlingkar diukur dan dibandingkan dengan masa aliran yang dijangkakan selama 35 minit. Didapati bahawa saluran bendalir dengan diameter 1.6 mm menunjukkan nilai yang paling hampir iaitu 34.33 minit (peratusan sisihan 1.91%) berbanding dua saluran lain. Reka bentuk saluran bendalir yang dioptimumkan (diameter saluran 1.6 mm) digabungkan dengan prestasi pemanasan pantas elemen PTC 230 (mencapai 65°C dalam 54.78 saat) untuk peranti LAMP mudah alih mewakili langkah signifikan ke arah pembangunan alat diagnostik POC yang pantas, tepat, dan mudah alih