13474 research outputs found
Sort by
Vehicle detection method based on optimised YOLOV7 lightweight model / Johan Lela Andika Johan Budiman
No full textThe advancement of unmanned aerial vehicles (UAVs) has encouraged researchers to update object detection algorithms for better accuracy and computational performance. Previous works that apply deep learning models for object detection applications required high graphics processing unit (GPU) computational power. Generally, object detection models suffer trade-off between accuracy and model size where the relationship is not always linear in deep learning models. Various factors such as architectural design, optimization techniques, and dataset characteristics can significantly influence the accuracy, model size and computational cost in adopting object detection models for low-cost embedded devices. Hence, it is crucial to employ lightweight object detection models for real-time object identification for the solution to be sustainable. This work proposes modifications on the head and backbone architecture of YOLOv7-tiny model. Firstly, efficient long-range aggregation network for vehicle detection (ELAN-VD) is incorporated in backbone layer. Secondly, the (UPSAMPLE-VD) on head architecture is improvised resolution to improve the detection accuracy of small vehicles in the aerial image. This study shows that the proposed method yields mean average precision (mAP) of 77.47 %, which is higher than the conventional YOLOv7-tiny of 48.89 %. In addition, the proposed model shown significant performance when compared to previous works, making it viable for application in low-cost embedded devices
Influence of digital literacy in classroom and attitudes on inclusive digital literacy pedagogy of in-service teachers in Malaysia / Wu Miaomiao
No full textThe convergence of digitization and inclusive education in IR 4.0 has instigated
alterations in educational content and teaching methods, necessitating educators to
possess digital literacy and adapt to novel pedagogical approaches, especially for
primary school teachers at the basic education level. Despite Malaysia’ s active
involvement in technological advancements and enhancements in inclusive education,
the outcomes have proven unsatisfactory. Therefore, this study identified the impact of
digital literacy on in-service teachers in inclusive classrooms and their attitudes toward
their inclusive digital literacy pedagogy in Malaysian primary schools during the IR 4.0
era. The study used cluster sampling and simple random sampling techniques to recruit
participants from Kuala Lumpur and Selangor, Malaysia. This study designed
instruments to measure teachers’ digital literacy, attitude, and inclusive digital literacy
pedagogy in inclusive classrooms. Reliability values for all three instruments were
above 0.9, factor loadings for each item were above 0.7, and model indices and
discriminant validity values were acceptable. The results indicated that the reliability
and validity of these instruments were satisfactory and can be used to assess the level of
teachers accordingly. The results of the assessment also revealed that teachers had a
positive attitude toward incorporating technology in inclusive classroom, possessed some knowledge of inclusive digital literacy pedagogy, but there were shortcomings in
action. Moreover, teachers’ digital literacy proficiency was moderate, particularly in
areas such as digital content creation and problem-solving skills, which remained
relatively underdeveloped. Regarding the model’ s design, PLS analysis was used to
assess and establish the model. Findings showed that teachers’ digital literacy in the
classroom has a significant impact on inclusive digital literacy pedagogy (β= 0.789,
T=27.895, P =0.000); teachers’ digital literacy in the classroom were positively related
to attitude (β= 0.776, T=26.991, P=0.000); teachers’ attitude have a direct impact on
their inclusive digital literacy pedagogy (β= 0.754, T=24. 188, P=0.000); and attitude
played a partially mediating role in the impact of teachers’ digital literacy in inclusive
classrooms on their inclusive digital literacy pedagogy. From the above, it can be
concluded that in IR 4.0, primary school teachers in inclusive education in Malaysia
also need to improve their digital literacy, continue to maintain and even enhance their
positive attitude towards integrating technology in an inclusive classroom, and endeavor
to know and do inclusive digital literacy pedagogy. An important way to realize this
vision is to emphasize and develop the professional development of teachers. This study
bridges the research gap on digital literacy in the classroom and inclusive digital literacy
pedagogy, which is innovative and academically valuable. Simultaneously, this study
also provides some references for in-service teachers’ future training. However, due to
objective constraints, this study has a limited study area and sample in Malaysia. Future
endeavors will focus on expanding and deepening research in this domain
Assessment of consistency detection in mandibular first premolars root canal morphology using U-NET model: Evaluation with dice coefficient and intersection over union – a pilot study / Melissa Wong Li Zheng
No full textThis study assesses the consistency of the U-Net model in pulp space segmentation of extracted mandibular first premolars (MFPs) using CBCT images. Five training, validation, and testing datasets were randomly generated from a pool of 130 CBCT images. Mean dice coefficient (DC) and Intersection over Union (IoU) scores were measured and compared across these datasets. The CBCT images were loaded into ITK-SNAP software (Version 4.0.2) for semi-automatic segmentation, performed by a postgraduate student (MW) with an ICC of 0.95. The datasets were split into five different training (70%), validation (20%), and testing (10%) sets from the same data pool. Image augmentation was applied, and the files were resized for the U-Net model. The mean DC and IoU scores for the training, validation, and testing datasets were compared using one-way ANOVA. This was followed by a post-hoc LSD test for multiple comparisons between the groups. A p-value < 0.05 was considered statistically significant. The mean DC from datasets 1 and 5 showed no significant difference during training, validation and testing phase with the p value of 0.32, 0.20, and 0.06 respectively. Similarly, the mean IoU from datasets 1 and 5 showed no significant difference during training, validation, and testing phase with the p value of 0.14, 0.23 and 0.12 respectively. In the training phase, the mean DC for dataset 2 (0.39 0.04) was significantly higher than dataset 1, 3, 4 and 5 (p = 0.00). Similarly, during the testing phase, the mean DC for dataset 2 (0.34 0.05) was significantly higher than dataset 1, 4 and 5 (p = 0.00), and dataset 3 (p = 0.02). The similar p-values for datasets 2 compared to datasets 1, 4 and 5 suggest that the performance differences between these datasets are consistent and significant in both the training and testing phases. The difference in p-values between datasets 2 and 3 during training and testing indicates that the performance gap was more pronounced during training than testing phases. The mean IoU for dataset 2 (0.40 0.07) performed significantly higher than dataset 1 (0.17 0.08, p = 0.00), 3 (0.26 0.10, p = 0.00), 4 (0.32 0.06, p = 0.04) and 5 (0.22 0.09, p = 0.00) during the training phase. Comparably, the mean IoU during the testing phase for dataset 2 (0.35 0.05) was significantly higher than dataset 1 (0.13 0.05, p = 0.00), 3 (0.27 0.07, p = 0.01), 4 (0.17 0.08, p = 0.00) and 5 (0.18 0.06, p = 0.00). The findings indicate that the consistency of the U-Net model in pulpal space segmentation of MFPs was affected despite the five datasets being randomly generated from the same data pool. When the predicted and ground truth images overlap completely, the DC and IoU value is 1. In the present study, the overall mean DC and IoU values across all phases were below the ideal value of 1, which indicates that the U-Net model performance in pulp segmentation is less consistent in the assessment of MFP
Identifying melanoma characteristics using directional imaging algorithm and convolutional neural network on dermoscopic images / Mohammad Asaduzzaman Rasel
No full textMelanoma is the deadliest skin cancer worldwide. Advancements in digital dermoscopic image analysis have greatly improved computer-aided Melanoma diagnosis systems. The use of dermoscopic images early detection of Melanoma has gained popularity among researchers due to its non-invasive nature. This thesis aims to enhance the analysis of dermoscopic images for identifying Melanoma. A critical first step for this is to distinguish between healthy and unhealthy skin areas by improving the segmentation process. This is followed by lesion features extraction and analysis (including classification) based on clinically diagnosis criteria including ABCDE rules, 3-point checklist, 7-point checklist, and CASH, to automate the manual process. This research is divided into two phases – 1) Feature Engineering phase explains skin conditions based on lesion segmentation and different dermoscopic feature extraction, while 2) Classification phase detects Melanoma. Multiple deep-learning models are proposed for segmentation. Several imaging, computer vision, and pattern recognition algorithms are employed to describe five dermoscopic features. Subsequently, these features are classified using different proposed deep learning models on various publicly available datasets. To overcome the issues with non-annotated dataset, several techniques are proposed. Both phases of the research outputs are evaluated and compared with the state-of-the-art methods. The proposed algorithms that outperformed the state-of-the-art algorithms contributes to diagnosing early-stage Melanoma. Findings from this study would help dermatologists and patients reduce the time and cost of Melanoma diagnosis, while receiving explanation for such automated diagnosis. Only five most common and important features of many Melanoma-features are analyzed. As part of future work, incorporating more dermoscopic features such as irregular blotches and regression structures in the analytical section would be promising
Optimizing fiber-reinforced lightweight high strength concrete: The role of coconut shells, ground granulated blast-furnace slag, and basalt fiber in low cement content mixes / Muhammed Talha Ünal
No full textThe demand for lightweight aggregate concrete (LWAC) has been growing in the
construction industry due to its potential to reduce the dead weight of structures while
maintaining strength, durability, and cost-effectiveness. This study proposes a method to
design LWAC by integrating coconut shell (CS) as coarse lightweight aggregate and a
high volume of wet-grinded ultrafine ground granulated blast furnace slag (UGGBS). The
wet grinding of GGBS has shown great potential for providing high-volume cement
replacement with reduced particle size and an improved hydration property while
simultaneously decreasing the density in slurry form. Also, to increase the effectiveness
of the hydration reactivity of UGGBS slurry, the ratios of GGBS powder and grinding
media volumes to the grinding chamber volume were examined. To optimize the mix
design of LWAC, a particle packing model was employed to estimate the packing density
and void ratio of the concrete mixture. The study has demonstrated that optimizing the
volume content of different CS and mining sand sizes can significantly improve packing
density. It is observed that higher packing densities of the aggregate phase and the entire
concrete mix lead to reduced void volumes, resulting in a less porous and more compacted
concrete matrix. Furthermore, the inclusion of fines and powder content has proven to be
an effective method for increasing packing density, which in turn enhances the voidfilling
capacity and strengthens the concrete. The optimal mix design was determined
using the packing density method, and the impact of Basalt Fiber (BF) was investigated
at varying levels (0%, 0.15%, and 1%). A comparative analysis was made between normal
weight concrete (NWC) and the selected LWAC mixtures with different BF contents in terms of cost, CO2 emissions, saturated surface dry density, surface crack conditions,
water absorption and porosity, sorptivity, compressive and flexural performances. The
results revealed that the incorporation of UGGBS had a substantial positive impact on the
mechanical properties of LWAC when BF was incorporated into the CS lightweight
concrete mixture. Including 1% BF in the mix demonstrated superior mechanical and
transport properties compared to non-fibrous LWAC, highlighting the significance of BF
in enhancing the characteristics of high-strength LWAC. As a significant finding of this
research, a grade 30 LWAC with a demoulded density of 1864 kg/m3 containing only 284
kg/m3 cement was produced. After considering all these approaches and implementations
of green high-strength LWAC with a %40 reduction in cement, there are lower CO2
emissions and reduced manufacturing costs, making it an alternative choice for
production
Risk factors associated with local recurrence in oral squamous cell carcinoma (OSCC) patients / Navenithamaria Eirutharajan
No full textBackground: Oral squamous cell carcinoma (OSCC) which is the most common histological subset of oral cancer remains a significant global health concern. The overall reported survival rate of OSCC is still poor and one of the main contributing factors for poor survival is local recurrence (LR) of the disease. Various clinicopathological factors are known to influence the LR of the OSCC. Objective: To identify the sociodemographic, clinicopathological characteristics and type of surgical interventions that influence the LR of OSCC. Methods: The socio-demographic, clinicopathological, and follow-up data of all the surgically treated OSCC cases from the year 2005 till September 2023 was obtained from the Diagnostic Oral Pathology Unit (DOPU) archive, which includes histopathological reports and H&E slides. Results: Out of 181 OSCC patients, 12 (6.6%) had LR. Of these, 8 were male and 4 females, with a mean age of 58. Malays had the highest LR percentage at 14.3%, followed by Chinese at 10% and Indians at 2.9%. Among 123 patients with risk habits, 9 (7.3%) had LR compared to 3 (5.2%) without habits. Primary tumours were mainly non-tongue regions, with 10 out of 12 LR cases present in the buccal mucosa, palate, and retromolar trigone. Among patients who underwent surgical resection only, 3 out of 40 (7.0%) developed LR. In contrast, among those who had both surgical resection and neck dissection, 9 out of 129 (6.5%) experienced LR. Bone invasion was present in 35 patients, with 6 (14.6%) experiencing LR. Multiple logistic regression identified race (p=0.024) and primary tumour site (p=0.020) as significant predictors of LR, with Malay and non-tongue tumours having a higher risk of developing LR. The model had an area under the ROC of 0.737, indicating moderate discrimination. Conclusion: Race and primary tumour site emerged as significant independent risk factors for LR. These findings highlight the need for targeted monitoring and tailored interventions to improve OSCC outcomes
Accurate identification of thirteen fly species from three families using wing venation patterns with machine learning approaches / Ling Min Hao
No full textThe ease and the affordability of image data acquisition have made whole-image analysis an attractive analytical approach in biological research. Coupled with machine learning, whole-image analysis has the potential to complement or even supplant traditional morphometric approaches for species identification in medical, veterinary, and forensic entomology. Here, I used a substantially expanded dataset (n = 759; 13 species and a species variant; 3 families) to consolidate findings from a pilot study (n = 74; 15 species; 2 families) for automated species identification of fly species based on their wing venation patterns, using classical Krawtchouk moment invariants coupled with a random forest model. To leverage on state-on-the-art methods on image analysis, I conducted a comparative analysis using ResNet, a deep learning model. Five-fold cross validation results show impressive mean identification accuracies of 98.56 ± 0.38% and 99.60 ± 0.27% at the family level, and 91.04 ± 1.33% and 97.87 ± 1.01% at the species level, for the classical and deep learning approaches, respectively. Additionally, the mean F1- scores of 0.89 ± 0.02 and 0.97 ± 0.01 respectively indicate a good balance of precision and recall for both models. Importantly, the regions on the fly wings that are used by ResNet for species identification were successfully visualised using Grad-CAM heatmaps, thus facilitating the interpretation of putative biological bases of identifications using ResNet. In summary, this study demonstrates the extent to which species differences in the studied dipteran species can be expressed in wing morphology, both quantitatively and qualitatively, through image data. Specifically, the findings from interpretable deep learning are potentially useful for generating hypotheses about putative wing anatomies that hold taxonomic value
The effect of Islamic basic principle on the Palestinian cause and its challenges: An analytical study in light of contemporary transformations / Aiman Ali Abboushi
No full textThe study examines the basic problem behind how the Palestinian Cause has grown in
Islamic, Arab, and humanitarian political and cultural consciousness. The study attempts
to confirm that Islamic constants have a real continuing impact on maintaining support
for the Cause through Islamic nature, especially after the decline of the leftist, Arab
nationalist, Palestinian nationalist, and international components. The study seeks to
determine the level of Arab, Islamic and humanitarian sympathy for the Palestinian
Cause, both official and popular. It builds on the Cause’s development from the
intellectual, doctrinal, and historical aspects from its inception until its formation in 1948,
to 1967 with the occupation of the rest of Palestine, through the peace process, global
transformations, and then the recent years that witnessed a decline in the Palestinian
Cause. The study asks how the Cause can’t be affected by the challenges it faces. It also
analyzes the stages of development of the Cause internationally, while studying the
motives and dimensions of its Islamic constants. It also raises the importance of the
emergence of these constants and their impact on the Cause from different angles,
especially in light of the political circumstances of the emergence and emergence of these
constants. The study uses a historical approach that includes analysis to trace the
relationship between historical Palestine and Islam in the Middle East. It is also based on
an analytical and legal investigative approach. The study aims to cover the most
prominent challenges facing Islamic constants that prevent them from fulfilling their
duty, and how to address these challenges in light of the regional and international reality
of the Palestinian Cause. It has become clear that the Palestinian Cause is facing a
conceptual conflict aimed at hollowing out its Islamic constants and changing its traditional terminology in order to marginalize its human rights dimension and turn it into
a regional crisis that can be overcome with an economic deal
Pricing and hedging exotic options in insurance and finance / Ng Ze-An
No full textThis thesis concerns the theoretical pricing and hedging of options, financial instruments
that give a payoff at a set date based on the price of one, or several other financial assets,
known as the underlying assets. The underlying assets are usually taken to be stocks, but
can also be bonds, securities, portfolios, or other financial instruments. In this thesis we
study two types of options - life contingent options and barrier Asian options. Because the
options examined in this thesis are relatively uncommon, with a novel mechanism of action,
they are known as exotic options. The analysis takes place in a stylised mathematical
model of a financial market known as the Black-Scholes model. We show for the life
contingent option that there exists a minimal super-hedging portfolio and determine the
associated initial investment. We also give a characterisation of when replication of the
option is possible. Next, we investigate the pricing problem for barrier Asian options with
short maturity times. Due to the nature of Asian options, closed form formulae for the fair
price of the option are relatively difficult to obtain. Using novel results from the theory
of stochastic calculus, we obtain closed form asymptotic formulae for the price of short
maturity barrier Asian options. Finally, we demonstrate our results with some explicit
examples
Characterization of secreted in Xylem (SIX) genes during plant- pathogen interaction of Fusarium oxysporum f.sp. cubense (Foc), Foc1 and Foc4 in Musa acuminata cv. ‘Berangan’ / Kausalyaa Kaliapan
No full textFusarium wilt of banana is a lethal disease caused by Fusarium oxysporum f.sp. cubense (Foc), a soil-borne pathogen that curbs the production of this crop. Small effector proteins known as Secreted in Xylem (SIX), are secreted into the xylem sap of the host plants triggering virulence and interferes with the defence responses of the host plants. Knowledge and comprehension of the interaction between Foc and banana plants during the infection is essential to expand methods to mitigate the disease’s damage. Therefore, studies on SIX genes are crucial as they have significant impact on Foc’s pathogenicity and play substantial role in the development of molecular framework and disease control approaches for banana. In this study, the responsible SIX genes for this pathogenic reaction were identified, characterized and their gene expression patterns were analysed during the infection. A comparative whole genome analysis was also carried out between Malaysian Foc1_C2HIR and Foc4_C1HIR isolates. A total of 48 cv. ‘Berangan’ plantlets were used where 32 plantlets were subjected to Foc infection (Foc1 and Foc4) while 16 plantlets were used as controls (without infection). Collection of plant samples were made at four different time points, 0 day, 2 weeks, 4 weeks and 5 weeks for molecular studies while the internal and external symptoms were confirmed by visual assessment and scored based on the percentage of infection on the 5th week post-inoculation. Following the data analysis, the ‘Berangan’ cultivar was confirmed to be highly susceptible to Foc infection. The Rhizome Discoloration Index (RDI) and Leaf Symptom Index (LSI) of Foc4-inoculated plants (7.25; 4.25 respectively) were higher compared to Foc1- inoculated plants (5.0; 3.25 respectively) which suggested that Foc4 is more virulent than Foc1 in cv. Berangan. Seven SIX genes i.e. SIX1, SIX2, SIX4, SIX6, SIX8a, SIX9a, SIX13 were identified to be present in Foc4 DNA, while only SIX9b was identified in Foc1 DNA and validated via sequencing analysis. Five of the SIX genes were shortlisted to be further studied on their expression pattern during infection by performing real-time PCR (qPCR) on fungal RNA from infected banana root samples. TEF1α and TUB were selected as the endogenous genes for normalisation. In the gene expression analysis conducted on Foc1 and Foc4 infected plants, two different patterns of expression were observed. In Foc4- inoculated plants, SIX1, SIX6, SIX8a and SIX9a were significantly upregulated on the 2nd and 5th week, whereas in Foc1-inoculated plants, significant upregulation of SIX9b was observed at a later stage which was on the 5th week. Besides, whole genome sequencing analysis for both Foc1_C2HIR and Foc4_C1HIR Malaysian isolates revealed that they had a respective genome size of 53 Mb and 55 Mb. Further analysis also revealed the genome features, SIX gene structures, putative virulence-associated genes, and functional groups of proteins identified in C2HIR and C1HIR isolates. These findings will aid in understanding the mechanism of potential disease control targets in these plants, including the improvement of diagnostics and breeding programmes