446494 research outputs found
Sort by
Comparative Analysis of Facial Recognition Based Attendance System Using Machine Learning
No full textAutomated attendance systems benefit from robust, scalable facial recognition models capable of handling real-world classroom variability illumination, pose, and occlusion. This study conducts a systematic comparison of four pipelines ArcFace, SFace, GhostFaceNet, and Dlib using the DeepFace and face recognition frameworks on a custom dataset of approximately 13,000 images representing 1,600 identities. Evaluation metrics include identification accuracy (Top-1), verification reliability (AUC, TAR@FAR), latency, and throughput (FPS). Dlib achieved the best overall performance (Top-1 = 87.93%, AUC = 0.9952, 11.53 FPS), outperforming deeper CNN-based embeddings in both speed and accuracy. The benchmark highlights trade-offs between accuracy, discriminability, and real-time efficiency, providing practical deployment insights for classroom automation
Passive Acoustic Detection and Localization of Drones Using MEMS Microphones and Machine Learning
No full textWith the rapid proliferation of drones in both civilian and military domains, the demand for efficient detection and tracking systems has become increasingly critical, particularly in sensitive and strategic areas. Conventional surveillance methods, such as radar and infrared sensing, often struggle to detect low-altitude, low-signature UAVs. This study proposes a real time acoustic localization system based on a distributed array of MEMS microphones. The approach utilizes Time Difference of Arrival (TDOA) estimations to determine the drone’s angular position, combined with a Random Forest classifier to distinguish drone acoustics from environmental noise. A radar-style interface was developed to provide real-time visualization of detections. Field experiments confirmed the system’s effectiveness under diverse environmental conditions. The solution offers a passive, cost-effective alternative for enhancing situational awareness in maritime and other security-sensitive applications
Probiotic Feed Technology Innovation for Food Security in the Fisheries Sector (Case of Tilapia Cultivation with Homemade Feed)
No full textFeed technology innovation is one of the key factors in supporting food security through the optimization of aquaculture productivity. Tilapia farming plays a strategic role in the development of the fisheries sector, with feed management being a determining factor in growth performance. This study aims to evaluate and compare the growth performance of tilapia (Oreochromis niloticus) fed with probiotic-containing pellets (Probio_FMUBB) and standard pellets without probiotics. The study was conducted from September to December 2022 in earthen ponds. Two floating net cages (10 x 10 meters) were used to house 1,500 tilapia fingerlings per cage with an average initial weight of 3.5 g. One cage was fed Probio_FMUBB, while the other served as the control. Samples were collected every two weeks from 10 fish per cage over five intervals. Results showed that tilapia fed with probiotic feed exhibited higher weight gain (61.6 g) compared to the control (54.5 g), although length gain was not significantly different (average 14.6 cm). Water quality parameters, such as temperature, pH, and phosphate, did not show significant differences, but ponds with probiotics had higher dissolved oxygen levels and lower ammonia levels. These findings support the potential use of probiotic feed to enhance tilapia farming productivity and promote sustainable food systems
Dining experience and customer satisfaction: Evidence from service quality and health-supportive food quality in Padang, Indonesia
No full textThis study examines how casual dining restaurants in Padang, Indonesia, respond to the growing demand for healthy eating by positioning health-supportive food quality and service quality as primary drivers of customer satisfaction, alongside atmosphere and physical environment as complementary factors. Data were collected from 123 customers through a structured questionnaire and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) to examine the relationships between dining experience attributes and satisfaction. The findings reveal that health-supportive food quality (β = 0.282, p = 0.001) and service quality (β = 0.472, p < 0.001) significantly enhance customer satisfaction, while atmosphere and physical environment do not have a significant impact. Descriptive analysis shows consistently high ratings for freshness, cleanliness, and serving temperature, but relatively lower ratings for the availability of highfiber, low-sugar menu options. These findings highlight the need for managers to expand health-oriented menu offerings and strengthen proactive service behaviors, while treating atmosphere and physical environment as baseline expectations
Determinants of Climate Change Perception Among Corn Farmers: A Structural Equation Modeling Analysis in Grobogan District, Indonesia
No full textThis study examines the determinants of climate change perception among corn farmers in Grobogan District, Central Java, one of Indonesia's main corn production areas that is increasingly exposed to climate risks. The research examines how individual characteristics, access to climate information, institutional support, and socioeconomic conditions affect farmers' perceptions, employing a quantitative explanatory approach. Data were collected from 340 purposively selected farmers and analyzed using Structural Equation Modelling-Partial Least Squares (SEM-PLS). The results indicate that all four exogenous variables significantly shape farmers' perceptions of climate change. Socioeconomic conditions emerge as the strongest determinant (β = 0.324; t = 6.095), followed by access to climate information (β = 0.292; t = 5.472), institutional support (β = 0.221; t = 4.121), and individual characteristics (β = 0.187; t = 3.732). The structural model demonstrates substantial explanatory power with an R2 of 0.639 and good predictive relevance (Q2 = 0.486). These findings underscore the importance of enhancing farmers' economic resilience, strengthening information dissemination, and improving institutional engagement in increasing awareness and readiness to respond to climate risks. The study provides empirical evidence for designing localized and multidimensional climate adaptation strategies to support smallholder farmers in climate-vulnerable regions
Development of an Antimalarial Recommender System to Accommodate Patient-specific Factors using an XGBoost Algorithm
No full textThe rate of malaria in Nigeria is alarmingly high however, what is more alarming is the quality of treatment of the most common disease in the country. Nigeria has the highest malaria morbidity rates in the world[1]. This study aimed at developing a chronic disease epidemiology risk system for recommending antimalarials in Nigeria based on patient-specific factors. This study explores the development of an antimalarial recommender system using an XGBoost algorithm. This algorithm was trained on a synthesized dataset, with a size of 516 rows, derived from the NAFDAC dataset containing the list of all endorsed antimalarials in Nigeria as well as the integration of the prescription guidelines given by WHO.Some of the factors considered when prescribing the drugs were weight, pregnancy status, malaria severity, and other medical condition. The chosen algorithm, XGBoost algorithm, was benchmarked against SVM, random forest, deep neural, and was selected due to its accuracy. From the XGBoost model, the accuracy result was 0.9811 and based on the System Usability Scale, the system scored a 83.15% highlighting how user-friendly the interface is. Overall, this antimalarial recommender system aids the quality of treatment of malaria
Synergistic Corrosion Inhibition of Mild Steel by Palmityltrimethylammonium Bromide and Methyl Protocatechualdehyde in Acidic Media: Experimental and Statistical Modeling Approaches
No full textThis study investigates the corrosion inhibition performance of a synergistic inhibitor system comprising palmityltrimethylammonium bromide and methyl protocatechualdehyde (PMB+MPA) on mild steel (MS) in 0.5 M H2SO4 and 0.5 M HCl solutions. The inhibition efficiency was assessed over 480 hours using seven concentrations ranging from 4.8 × 106 M to 2.9 × 105 M. Corrosion rates and inhibition efficiencies were determined through gravimetric analysis, and the effects of concentration and exposure time were statistically evaluated using ANOVA. Results show that inhibition efficiency increases with both exposure time and inhibitor concentration, reaching a plateau in H2SO4 (~95–97%) beyond 288 h due to surface saturation. In HCl, however, efficiency peaks earlier (~120–144 h) and declines significantly at lower concentrations due to Cl~ ion aggression. A second-order polynomial model was proposed to describe inhibition efficiency as a function of time and concentration. Overall, PMB+MPA demonstrated superior long-term stability in H2SO4, with notable concentration-dependent sensitivity in HCl, suggesting its suitability as an eco-friendly inhibitor in less aggressive environments