Search results for: GPU architectures

Authors: Irina Alexandra Paun, Mona Mihailescu, Marian Zamfirescu, Catalin Romeo Luculescu, Adriana Maria Acasandrei, Cosmin Catalin Mustaciosu, Roxana Cristina Popescu, Maria Dinescu

Abstract:

A major concern in bone tissue engineering is to develop complex 3D architectures that mimic the natural cells environment, facilitate the cells growth in a defined manner and allow the flow transport of nutrients and metabolic waste. In particular, porous structures of controlled pore size and positioning are indispensable for growing human-like bone structures. Another concern is to monitor both the structures and the seeded cells with high spatial resolution and without interfering with the cells natural environment. The present approach relies on laser-based technologies employed for fabricating porous biomimetic structures that support the growth of osteoblast-like cells and for their non-invasive 3D imaging. Specifically, the porous structures were built by two photon polymerization –direct writing (2PP_DW) of the commercially available photoresists IL-L780, using the Photonic Professional 3D lithography system. The structures consist of vertical tubes with micrometer-sized heights and diameters, in a honeycomb-like spatial arrangement. These were fabricated by irradiating the IP-L780 photoresist with focused laser pulses with wavelength centered at 780 nm, 120 fs pulse duration and 80 MHz repetition rate. The samples were precisely scanned in 3D by piezo stages. The coarse positioning was done by XY motorized stages. The scanning path was programmed through a writing language (GWL) script developed by Nanoscribe. Following laser irradiation, the unexposed regions of the photoresist were washed out by immersing the samples in the Propylene Glycol Monomethyl Ether Acetate (PGMEA). The porous structures were seeded with osteoblast like MG-63 cells and their osteogenic potential was tested in vitro. The cell-seeded structures were analyzed in 3D using the digital holographic microscopy technique (DHM). DHM is a marker free and high spatial resolution imaging tool, where the hologram acquisition is performed non-invasively i.e. without interfering with the cells natural environment. Following hologram recording, a digital algorithm provided a 3D image of the sample, as well as information about its refractive index, which is correlated with the intracellular content. The axial resolution of the images went down to the nanoscale, while the temporal scales ranged from milliseconds up to hours. The hologram did not involve sample scanning and the whole image was available in one frame recorded going over 200μm field of view. The digital holograms processing provided 3D quantitative information on the porous structures and allowed a quantitative analysis of the cellular response in respect to the porous architectures. The cellular shape and dimensions were found to be influenced by the underlying micro relief. Furthermore, the intracellular content gave evidence on the beneficial role of the porous structures in promoting osteoblast differentiation. In all, the proposed laser-based protocol emerges as a promising tool for the fabrication and non-invasive imaging of porous constructs for bone tissue engineering. Acknowledgments: This work was supported by a grant of the Romanian Authority for Scientific Research and Innovation, CNCS-UEFISCDI, project PN-II-RU-TE-2014-4-2534 (contract 97 from 01/10/2015) and by UEFISCDI PN-II-PT-PCCA no. 6/2012. A part of this work was performed in the CETAL laser facility, supported by the National Program PN 16 47 - LAPLAS IV.

Keywords: biomimetic, holography, laser, osteoblast, two photon polymerization

Procedia PDF Downloads 271

Authors: Olanrewaju B. Lasisi

Abstract:

The position of Ijebu-Ode, the historical capital of the Ijebu Kingdom, at the center of gravity of Ijebu land is enclosed by the 180-km-long earthwork and suggests a centrally controlled project. This paper reflects on the first stratigraphic drawing of the banks and ditches of this earthwork, and place its construction mechanism in a chronological framework. Nine radiocarbon dates obtained at the site suggest that the earthwork was built in the late 14th or early 15th century. This suggests a relationship with the Ijebu Kingdom, which pre-existed the opening of the Atlantic trade but first became visible only in the Portuguese records in the 1480s. In June 2017, more earthworks were found but within the core of Ijebu Land. This most recent finding points to an extension of territory from the center to the outlying villages. One central question about this discovery of monumental architectures that was functional around the 14th century or before is in its mode of construction. Apparently, iron tools must have been used in the construction of ‘a 20m deep ditch that runs 180km in circumference.’ Thus, the discovery of iron-working sites around the vicinity of the earthwork is a pointer to this building process that is up till now shrouded in mystery. By comparing the chronology of Ijebu earthworks with the evidence of Iron working in south western Nigeria around the first half of the first millennium AD, it can be thought that the rise in polity triggered the knowledge of metallurgy in the region.

Keywords: archaeology, earthworks, Ijebu, metallurgy

Procedia PDF Downloads 243

Authors: Bunakiye R. Japheth, Ogude U. Cyril

Abstract:

Since the process of transforming user requirements to modeling constructs are not very well supported by domain-specific frameworks, it became necessary to integrate domain requirements with the specific architectures to achieve an integrated customizable solutions space via artifact orientation. Domain-specific modeling language specifications of model-driven engineering technologies focus more on requirements within a particular domain, which can be tailored to aid the domain expert in expressing domain concepts effectively. Modeling processes through domain-specific language formalisms are highly volatile due to dependencies on domain concepts or used process models. A capable solution is given by artifact orientation that stresses on the results rather than expressing a strict dependence on complicated platforms for model creation and development. Based on this premise, domain-specific methods for producing artifacts without having to take into account the complexity and variability of platforms for model definitions can be integrated to support customizable development. In this paper, we discuss methods for the integration capabilities and necessities within a common structure and semantics that contribute a metamodel for artifact-orientation, which leads to a reusable software layer with concrete syntax capable of determining design intents from domain expert. These concepts forming the language formalism are established from models explained within the oil and gas pipelines industry.

Keywords: control process, metrics of engineering, structured abstraction, semantic model

Procedia PDF Downloads 141

Authors: Xin Zhao

Abstract:

3D printing is used in creating bone grafts of various architectures by printing materials in a layer-by-layer manner. Traditionally, to make materials printable, heating up or dissolving materials in organic solvents have been used, compromising their capability in loading biomolecules. Photocrosslinkable materials which are initially liquid and printable, and solidified upon light exposure are therefore developed. However, the existing photocrosslinkable materials are either too soft to bear load or non-degradable with potential long-term biocompatibility problems. Here, photocrosslinkable nanocomposite ink is developed composed of poly (lactide-co-propylene glycol-co-lactide) dimethacrylate (PmLnDMA) and hydroxyethyl methacrylate-functionalized hydroxyapatite nanoparticles (nHAMA) mimicking the hairy setae of gecko that can strongly interact with its surroundings to bear high load. Incorporation of nHAMA into PmLnDMA endows the nanocomposite ink with several advantages in (1) improved organic/inorganic interfacial compatibility to increase mechanical strength, (2) readily modulated rheological behaviors, wettability, and biodegradation, (3) enhanced osteoconductivity and osteoinductivity. Moreover, the ink can be rapidly crosslinked upon light exposure, load, and long-term release growth factors, and be printed into 3D bone scaffolds of various shapes and structures according to the patients’ needs. Altogether, this innovation will benefit patients all over the world who suffer from bone fractures, tumors, infections.

Keywords: photocrosslinkable nanocomposite, 3D printing, bone ink, personalized medicine

Procedia PDF Downloads 114

Authors: Rujia Chen, Ajit Narayanan

Abstract:

Convolutional neural networks (CNN), as typically applied in deep learning, use layer-wise backpropagation (BP) to construct filters and kernels for feature extraction. Such filters are 2D or 3D groups of weights for constructing feature maps at subsequent layers of the CNN and are shared across the entire input. BP as a gradient descent algorithm has well-known problems of getting stuck at local optima. The use of genetic algorithms (GAs) for evolving weights between layers of standard artificial neural networks (ANNs) is a well-established area of neuroevolution. In particular, the use of crossover techniques when optimizing weights can help to overcome problems of local optima. However, the application of GAs for evolving the weights of filters and kernels in CNNs is not yet an established area of neuroevolution. In this paper, a GA-based filter development algorithm is proposed. The results of the proof-of-concept experiments described in this paper show the proposed GA algorithm can find filter weights through evolutionary techniques rather than BP learning. For some simple classification tasks like geometric shape recognition, the proposed algorithm can achieve 100% accuracy. The results for MNIST classification, while not as good as possible through standard filter learning through BP, show that filter and kernel evolution warrants further investigation as a new subarea of neuroevolution for deep architectures.

Keywords: neuroevolution, convolutional neural network, genetic algorithm, filters, kernels

Procedia PDF Downloads 185

Authors: Yulong Jia, Xiang Zhang, Ziyuan Wu, Shiji Hu

Abstract:

The calculation of bilinear pairing is the core of the SM9 algorithm, which relies on the underlying prime domain algorithm and the quadratic extension domain algorithm. Among the field algorithms, modular multiplication operation is the most time-consuming part. Therefore, the underlying modular multiplication algorithm is optimized to maximize the operation speed of bilinear pairings. This paper uses a modular multiplication method based on non-least positive (NLP) combined with Karatsuba and schoolbook multiplication to improve the Montgomery algorithm. At the same time, according to the characteristics of multiplication operation in the quadratic extension domain, a quadratic extension domain FP2-NLP modular multiplication algorithm for bilinear pairings is proposed, which effectively reduces the operation time of modular multiplication in the quadratic extension domain. The sub-expanded domain Fp₂ -NLP modular multiplication algorithm effectively reduces the operation time of modular multiplication under the second-expanded domain. The multiplication unit in the quadratic extension domain is implemented using SMIC55nm process, and two different implementation architectures are designed to cope with different application scenarios. Compared with the existing related literature, The output latency of this design can reach a minimum of 15 cycles. The shortest time for calculating the (AB+CD)r⁻¹ mod form is 37.5ns, and the comprehensive area-time product (AT) is 11400. The final R-ate pairing algorithm hardware accelerator consumes 2670k equivalent logic gates and 1.8ms computing time in 55nm process.

Keywords: sm9, hardware, NLP, Montgomery

Procedia PDF Downloads 0

Authors: Marrone Silverio Melo Dantas Pedro Henrique Dreyer, Gabriel Fonseca Reis de Souza, Daniel Bezerra, Ricardo Souza, Silvia Lins, Judith Kelner, Djamel Fawzi Hadj Sadok

Abstract:

The creation of a dataset is time-consuming and often discourages researchers from pursuing their goals. To overcome this problem, we present and discuss two solutions adopted for the automation of this process. Both optimize valuable user time and resources and support video object segmentation with object tracking and 3D projection. In our scenario, we acquire images from a moving robotic arm and, for each approach, generate distinct annotated datasets. We evaluated the precision of the annotations by comparing these with a manually annotated dataset, as well as the efficiency in the context of detection and classification problems. For detection support, we used YOLO and obtained for the projection dataset an F1-Score, accuracy, and mAP values of 0.846, 0.924, and 0.875, respectively. Concerning the tracking dataset, we achieved an F1-Score of 0.861, an accuracy of 0.932, whereas mAP reached 0.894. In order to evaluate the quality of the annotated images used for classification problems, we employed deep learning architectures. We adopted metrics accuracy and F1-Score, for VGG, DenseNet, MobileNet, Inception, and ResNet. The VGG architecture outperformed the others for both projection and tracking datasets. It reached an accuracy and F1-score of 0.997 and 0.993, respectively. Similarly, for the tracking dataset, it achieved an accuracy of 0.991 and an F1-Score of 0.981.

Keywords: RJ45, automatic annotation, object tracking, 3D projection

Procedia PDF Downloads 166

Authors: Azeez Shaik, Amit Salvi, B. P. Gautham

Abstract:

Fiber reinforced polymer resin composite materials are finding wide variety of applications in automotive and aerospace industry because of their high specific stiffness and specific strengths when compared to metals. New class of 2D and 3D textile and woven fabric composites offer excellent fracture toughens as they bridge the cracks formed during fracture. Due to complexity of their fiber architectures and its resulting composite microstructures, optimized design and analysis of these structures is very complicated. A traditional homogenization approach is typically used to analyze structures made up of these materials. This approach usually fails to predict damage initiation as well as damage propagation and ultimate failure of structure made up of woven and textile composites. This study demonstrates a methodology to analyze woven and textile composites by using the multi-level multi-scale modelling approach. In this approach, a geometric repetitive unit cell (RUC) is developed with all its constituents to develop a representative volume element (RVE) with all its constituents and their interaction modeled correctly. The structure is modeled based on the RUC/RVE and analyzed at different length scales with desired levels of fidelity incorporating the damage and failure. The results are passed across (up and down) the scales qualitatively as well as quantitatively from the perspective of material, configuration and architecture.

Keywords: cohesive zone, multi-scale modeling, rate dependency, RUC, woven textiles

Procedia PDF Downloads 360

Authors: Sekkal Nawel

Abstract:

The Web of Things (WoT), integration of smart technologies from the Internet or network to Web architecture or application, is becoming more complex, larger, and dynamic. The WoT is associated with various elements such as sensors, devices, networks, protocols, data, functionalities, and architectures to perform services for stakeholders. These services operate in the context of the interaction of stakeholders and the WoT elements. Such context is becoming a key information source from which data are of various nature and uncertain, thus leading to complex situations. In this paper, we take interest in the development of intelligent Web services. The key ingredients of this “intelligent” notion are the context diversity, the necessity of a semantic representation to manage complex situations and the capacity to reason with uncertain data. In this perspective, we introduce a multi-layered architecture based on a generic intelligent Web service model dealing with various contexts, which proactively predict future situations and reactively respond to real-time situations in order to support decision-making. For semantic context data representation, we use PR-OWL, which is a probabilistic ontology based on Multi-Entity Bayesian Networks (MEBN). PR-OWL is flexible enough to represent complex, dynamic, and uncertain contexts, the key requirements of the development for the intelligent Web services. A case study was carried out using the proposed architecture for intelligent plant watering to show the role of proactive and reactive contextual reasoning in terms of WoT.

Keywords: smart web service, the web of things, context reasoning, proactive, reactive, multi-entity bayesian networks, PR-OWL

Procedia PDF Downloads 70

Authors: Amrutha Jayan

Abstract:

The cities are attractors of the human population, offering opportunities for economic activities for different linguistic, cultural, and ethnic groups. The urban form and design of the city impact the life of these people. Social and cultural exclusions result in spatial segregation and gentrification. The spaces provided in cities must be inclusive for all these communities for them to feel part of the city and contribute to society. Intercultural urbanism is a theory and practice of city building, planning, and design of urban spaces and architectures that are cognizant of the social impact of the built environment. The postulate acknowledges cultural differences and opportunities for cultural exchange. Literature on intercultural urbanism, culture and space, spatial justice, and cultural inclusion are analyzed to identify parameters contributing to intercultural placemaking. A qualitative study on Mattancherry shows how the precinct has sustained throughout the years with different communities living together within a radius of 5 km, creating a diverse and vibrant environment. The research identifies the urban elements that contribute to intercultural interactions and maintain the synergy between these communities. The public spaces, porous edges, built-form, streets, and accessibility contribute to chance encounters and intercultural interactivity. The research seeks to find the factors that contribute to intercultural placemaking.

Keywords: intercultural urbanism, cultural inclusion, spatial justice, public space

Procedia PDF Downloads 219

Authors: Jie Zhao, Meng Su

Abstract:

Image recognition, as one of the most critical technologies in computer vision, works to help machine-like robotics understand a scene, that is, if deployed appropriately, will trigger the revolution in remote sensing and industry automation. With the developments of AI technologies, there are many prevailing and sophisticated neural networks as technologies developed for image recognition. However, computer vision platforms as hardware, supporting neural networks for image recognition, as crucial as the neural network technologies, need to be more congruently addressed as the research subjects. In contrast, different computer vision platforms are deterministic to leverage the performance of different neural networks for recognition. In this paper, three different computer vision platforms – Jetson Nano(with 4GB), a standalone laptop(with RTX 3000s, using CUDA), and Google Colab (web-based, using GPU) are explored and four prominent neural network architectures (including AlexNet, VGG(16/19), GoogleNet, and ResNet(18/34/50)), are investigated. In the context of pairwise usage between different computer vision platforms and distinctive neural networks, with the merits of recognition accuracy and time efficiency, the performances are evaluated. In the case study using public imageNets, our findings provide a nuanced perspective on optimizing image recognition tasks across Edge-AI platforms, offering guidance on selecting appropriate neural network structures to maximize performance under hardware constraints.

Keywords: alexNet, VGG, googleNet, resNet, Jetson nano, CUDA, COCO-NET, cifar10, imageNet large scale visual recognition challenge (ILSVRC), google colab

Procedia PDF Downloads 89

Authors: A. Degale Desta, Cheng Jian

Abstract:

Deep learning application in computer vision is rapidly advancing, giving it the ability to monitor the public and quickly identify potentially anomalous behaviour from crowd scenes. Therefore, the purpose of the current work is to improve the performance of safety of people in crowd events from panic behaviour through introducing the innovative idea of Aggregation of Ensembles (AOE), which makes use of the pre-trained ConvNets and a pool of classifiers to find anomalies in video data with packed scenes. According to the theory of algorithms that applied K-means, KNN, CNN, SVD, and Faster-CNN, YOLOv5 architectures learn different levels of semantic representation from crowd videos; the proposed approach leverages an ensemble of various fine-tuned convolutional neural networks (CNN), allowing for the extraction of enriched feature sets. In addition to the above algorithms, a long short-term memory neural network to forecast future feature values and a handmade feature that takes into consideration the peculiarities of the crowd to understand human behavior. On well-known datasets of panic situations, experiments are run to assess the effectiveness and precision of the suggested method. Results reveal that, compared to state-of-the-art methodologies, the system produces better and more promising results in terms of accuracy and processing speed.

Keywords: action recognition, computer vision, crowd detecting and tracking, deep learning

Procedia PDF Downloads 161

Authors: Sifatullah Bahij, Safiullah Omary, Francoise Feugeas, Amanullah Faqiri

Abstract:

Reinforced concrete (RC) is commonly used material in the construction sector, due to its low-cost and durability, and allowed the architectures and designers to construct structural members with different shapes and finishing. Usually, RC members are designed to sustain service loads efficiently without any destruction. However, because of the faults in the design phase, overloading, materials deficiencies, and environmental effects, most of the structural elements will require maintenance and repairing over their lifetime. Therefore, strengthening and repair of the deteriorated and/or existing RC structures are much important to extend their life cycle. Various techniques are existing to retrofit and strengthen RC structural elements such as steel plate bonding, external pre-stressing, section enlargement, fiber reinforced polymer (FRP) wrapping, etc. Although these configurations can successfully improve the load bearing capacity of the beams, they are still prone to corrosion damage which results in failure of the strengthened elements. Therefore, many researchers used fiber reinforced cementitious materials due to its low-cost, corrosion resistance, and result in improvement of the tensile and fatigue behaviors. Various types of cementitious materials have been used to strengthen or repair structural elements. This paper has summarized to accumulate data regarding on previously published research papers concerning the torsional behaviors of RC beams strengthened by various types of cementitious materials.

Keywords: reinforced concrete beams, strengthening techniques, cementitious materials, torsional strength, twisting angle

Procedia PDF Downloads 119

Authors: Ferdinando Montemari, Antonio Vitale, Nicola Genito, Luca Garbarino, Urbano Tancredi, Domenico Accardo, Michele Grassi, Giancarmine Fasano, Anna Elena Tirri

Abstract:

The development of guidance, navigation and control algorithms and avionic procedures requires the disposability of suitable analysis and verification tools, such as simulation environments, which support the design process and allow detecting potential problems prior to the flight test, in order to make new technologies available at reduced cost, time and risk. This paper presents a simulation environment for avionic software development and qualification, especially aimed at equipment for general aviation aircrafts and unmanned aerial systems. The simulation environment includes models for short and medium-range radio-navigation aids, flight assistance systems, and ground control stations. All the software modules are able to simulate the modeled systems both in fast-time and real-time tests, and were implemented following component oriented modeling techniques and requirement based approach. The paper describes the specific models features, the architectures of the implemented software systems and its validation process. Performed validation tests highlighted the capability of the simulation environment to guarantee in real-time the required functionalities and performance of the simulated avionics systems, as well as to reproduce the interaction between these systems, thus permitting a realistic and reliable simulation of a complete mission scenario.

Keywords: ADS-B, avionics, NAVAIDs, real-time simulation, TCAS, UAS ground control station

Procedia PDF Downloads 226

Authors: Muhammad Mohsin, Fawad Ahmad, Fatima Batool, Muhammad Kaab Zarrar

Abstract:

The integration of the Internet of Things (IoT) and edge computing devices offers a transformative approach to electronic waste management, particularly in the dismantling of printed circuit boards (PCBs). This paper explores how these technologies optimize operational efficiency and improve environmental sustainability by addressing challenges such as data security, interoperability, scalability, and real-time data processing. Proposed solutions include advanced machine learning algorithms for predictive maintenance, robust encryption protocols, and scalable architectures that incorporate edge computing. Case studies from leading e-waste management facilities illustrate benefits such as improved material recovery efficiency, reduced environmental impact, improved worker safety, and optimized resource utilization. The findings highlight the potential of IoT and edge computing to revolutionize e-waste dismantling and make the case for a collaborative approach between policymakers, waste management professionals, and technology developers. This research provides important insights into the use of IoT and edge computing to make significant progress in the sustainable management of electronic waste

Keywords: internet of Things, edge computing, waste PCB disassembly, electronic waste management, data security, interoperability, machine learning, predictive maintenance, sustainable development

Procedia PDF Downloads 29

Authors: Ayse Cagil Kandemir, Derya Erdem, Markus Niederberger, Ralph Spolenak

Abstract:

Dip Pen nanolithography (DPN), which is a tip based method, serves a novel approach to produce nano and micro-scaled patterns due to its high resolution and pattern flexibility. It is introduced as a new constructive scanning probe lithography (SPL) technique. DPN delivers materials in the form of an ink by using the tip of a cantilever as pen and substrate as paper in order to form surface architectures. First studies rely on delivery of small organic molecules on gold substrate in ambient conditions. As time passes different inks such as; polymers, colloidal particles, oligonucleotides, metallic salts were examined on a variety of surfaces. Discovery of DPN also enabled patterning with multiple inks by using multiple cantilevers for the first time in SPL history. Specifically, polymer inks, which constitute a flexible matrix for various materials, can have a potential in MEMS, NEMS and drug delivery applications. In our study, it is aimed to construct polymer patterns using DPN by studying wetting behavior of polymer on semiconductor, metal and polymer surfaces. The optimum viscosity range of polymer and effect of environmental conditions such as humidity and temperature are examined. It is observed that there is an inverse relation with ink viscosity and depletion time. This study also yields the optimal writing conditions to produce consistent patterns with DPN. It is shown that written dot sizes increase with dwell time, indicating that the examined writing conditions yield repeatable patterns.

Keywords: dip pen nanolithography, polymer, surface patterning, surface science

Procedia PDF Downloads 396

Authors: Seon Tae Kim, Robert Hegner, Goksel Ozuylasi, Panagiotis Stathopoulos, Eberhard Nicke

Abstract:

High-temperature heat pumps (HTHP) that can supply heat at temperatures above 200°C can enhance the energy efficiency of industrial processes and reduce the CO₂ emissions connected with the heat supply of these processes. In the current work, the thermodynamic performance of 3 different vapor compression cycles, which use R-718 (water) as a working medium, have been evaluated by using a commercial process simulation tool (EBSILON Professional). All considered cycles use two-stage vapor compression with intercooling between stages. The main aim of the study is to compare different intercooling strategies and study possible heat recovery scenarios within the intercooling process. This comparison has been carried out by computing the coefficient of performance (COP), the heat supply temperature level, and the respective mass flow rate of water for all cycle architectures. With increasing temperature difference between the heat source and heat sink, ∆T, the COP values decreased as expected, and the highest COP value was found for the cycle configurations where both compressors have the same pressure ratio (PR). The investigation on the HTHP capacities with optimized PR and exergy analysis has also been carried out. The internal heat exchanger cycle with the inward direction of secondary flow (IHX-in) showed a higher temperature level and exergy efficiency compared to other cycles. Moreover, the available operating range was estimated by considering mechanical limitations.

Keywords: high temperature heat pump, industrial process, vapor compression cycle, R-718 (water), thermodynamic analysis

Procedia PDF Downloads 147

Authors: Simone Leonardi, Luca Ardito

Abstract:

Micro-influencers are effective elements in the marketing strategies of companies and institutions because of their capability to create an hyper-engaged audience around a specific topic of interest. In recent years, many scientific approaches and commercial tools have handled the task of detecting this type of social media users. These strategies adopt solutions ranging from rule based machine learning models to deep neural networks and graph analysis on text, images, and account information. This work compares the existing solutions and proposes an ensemble method to generalize them with different input data and social media platforms. The deployed solution combines deep learning models on unstructured data with statistical machine learning models on structured data. We retrieve both social media accounts information and multimedia posts on Twitter and Instagram. These data are mapped into feature vectors for an eXtreme Gradient Boosting (XGBoost) classifier. Sixty different topics have been analyzed to build a rule based gold standard dataset and to compare the performances of our approach against baseline classifiers. We prove the effectiveness of our work by comparing the accuracy, precision, recall, and f1 score of our model with different configurations and architectures. We obtained an accuracy of 0.91 with our best performing model.

Keywords: deep learning, gradient boosting, image processing, micro-influencers, NLP, social media

Procedia PDF Downloads 183

Authors: Seren Sevim Öğmen, Ömer Özyiğit

Abstract:

Vouni Palace is a Persian palace built on a rocky hill in the Lefke district of Cyprus. The palace is one of the limited number of architectures identified, which prove the existence of a Persian period on the island. Since the excavations on the palace were held a very long time ago, there is a need to re-date the cultural layers within the palace using new archaeological evidence and recent studies. The existing chronology has been reviewed and a new chronology has been created according to its architectural structure, floor findings such as ceramics and sculptures and the stratigraphic layer of Room 59 where the Vouni Treasure was found. This work dates the palace in Vouni between the periods of c. 520 BC, deduced from the early period sculptures, and c. 330 BC by the late period floor ceramics. Some earlier dated archaic sculptures are identified in Room 122 – which takes part in the temenos area of the palace, and correspondingly the construction of the palace is dated c. 520 BC. The comparison between Vouni Palace and Persian palaces built in Iran, shows similarities with palaces built during the rule of Darius. It is evident that two main building periods of the palace which are previously identified, represent Persian influence according to its architectural structure and findings. Several floor potteries show that there must be other layer or layers after Vouni Treasure dated 390/380 BC, which was considered as the destruction date of the palace. At this point the forenamed date can indicate the end of a stage, not the end of the period because the palace was still in use until c. 330 BC. The results of the study, in addition to dating the layers of Vouni Palace, enlightens the administrative function of the Palace within the Persian rule in Cyprus.

Keywords: administrative, chronology, cyprus, persian rule, vouni palace

Procedia PDF Downloads 117

Authors: Ali Almadan, Anoop Krishnan, Ajita Rattani

Abstract:

Facial recognition is a widely adopted technology in surveillance, border control, healthcare, banking services, and lately, in mobile user authentication with Apple introducing “Face ID” moniker with iPhone X. A lot of research has been conducted in the area of face recognition on datasets captured by surveillance cameras, DSLR, and mobile devices. Recently, face recognition technology has also been deployed on body-worn cameras to keep officers safe, enabling situational awareness and providing evidence for trial. However, limited academic research has been conducted on this topic so far, without the availability of any publicly available datasets with a sufficient sample size. This paper aims to advance research in the area of face recognition using body-worn cameras. To this aim, the contribution of this work is two-fold: (1) collection of a dataset consisting of a total of 136,939 facial images of 102 subjects captured using body-worn cameras in in-door and daylight conditions and (2) evaluation of various deep-learning architectures for face identification on the collected dataset. Experimental results suggest a maximum True Positive Rate(TPR) of 99.86% at False Positive Rate(FPR) of 0.000 obtained by SphereFace based deep learning architecture in daylight condition. The collected dataset and the baseline algorithms will promote further research and development. A downloadable link of the dataset and the algorithms is available by contacting the authors.

Keywords: face recognition, body-worn cameras, deep learning, person identification

Procedia PDF Downloads 161

Authors: M. K. Benhaoua, A. K. Singh, A. E. Benyamina, P. Boulet

Abstract:

Mapping parallelized tasks of applications onto these MPSoCs can be done either at design time (static) or at run-time (dynamic). Static mapping strategies find the best placement of tasks at design-time, and hence, these are not suitable for dynamic workload and seem incapable of runtime resource management. The number of tasks or applications executing in MPSoC platform can exceed the available resources, requiring efficient run-time mapping strategies to meet these constraints. This paper describes a new Spiral Dynamic Task Mapping heuristic for mapping applications onto NoC-based Heterogeneous MPSoC. This heuristic is based on packing strategy and routing Algorithm proposed also in this paper. Heuristic try to map the tasks of an application in a clustering region to reduce the communication overhead between the communicating tasks. The heuristic proposed in this paper attempts to map the tasks of an application that are most related to each other in a spiral manner and to find the best possible path load that minimizes the communication overhead. In this context, we have realized a simulation environment for experimental evaluations to map applications with varying number of tasks onto an 8x8 NoC-based Heterogeneous MPSoCs platform, we demonstrate that the new mapping heuristics with the new modified dijkstra routing algorithm proposed are capable of reducing the total execution time and energy consumption of applications when compared to state-of-the-art run-time mapping heuristics reported in the literature.

Keywords: multiprocessor system on chip, MPSoC, network on chip, NoC, heterogeneous architectures, run-time mapping heuristics, routing algorithm

Procedia PDF Downloads 489

Authors: Smaran Manchala

Abstract:

Homomorphic Encryption (HE) enables computations on encrypted data without requiring decryption, mitigating data vulnerability during processing. Usable Fully Homomorphic Encryption (FHE) could revolutionize secure data operations across cloud computing, AI training, and healthcare, providing both privacy and functionality, however, the computational inefficiency of schemes like Cheon-Kim-Kim-Song (CKKS) hinders their widespread practical use. This study focuses on optimizing CKKS for faster matrix operations through the implementation of vector computation parallelization and GPU acceleration. The variable effects of vector parallelization on GPUs were explored, recognizing that while parallelization typically accelerates operations, it could introduce overhead that results in slower runtimes, especially in smaller, less computationally demanding operations. To assess performance, two neural network models, MLPN and CNN—were tested on the MNIST dataset using both ARM and x86-64 architectures, with CNN chosen for its higher computational demands. Each test was repeated 1,000 times, and outliers were removed via Z-score analysis to measure the effect of vector parallelization on CKKS performance. Model accuracy was also evaluated under CKKS encryption to ensure optimizations did not compromise results. According to the results of the trail runs, applying vector parallelization had a 2.63X efficiency increase overall with a 1.83X performance increase for x86-64 over ARM architecture. Overall, these results suggest that the application of vector parallelization in tandem with GPU acceleration significantly improves the efficiency of CKKS even while accounting for vector parallelization overhead, providing impact in future zero trust operations.

Keywords: CKKS scheme, runtime efficiency, fully homomorphic encryption (FHE), GPU acceleration, vector parallelization

Procedia PDF Downloads 23

Authors: Maria Amodeo, Pasquale Arpaia, Marco Buzio, Vincenzo Di Capua, Francesco Donnarumma

Abstract:

Different deep neural network architectures have been compared and tested to predict magnetic hysteresis in the context of pulsed electromagnets for experimental physics applications. Modelling quasi-static or dynamic major and especially minor hysteresis loops is one of the most challenging topics for computational magnetism. Recent attempts at mathematical prediction in this context using Preisach models could not attain better than percent-level accuracy. Hence, this work explores neural network approaches and shows that the architecture that best fits the measured magnetic field behaviour, including the effects of hysteresis and eddy currents, is the nonlinear autoregressive exogenous neural network (NARX) model. This architecture aims to achieve a relative RMSE of the order of a few 100 ppm for complex magnetic field cycling, including arbitrary sequences of pseudo-random high field and low field cycles. The NARX-based architecture is compared with the state-of-the-art, showing better performance than the classical operator-based and differential models, and is tested on a reference quadrupole magnetic lens used for CERN particle beams, chosen as a case study. The training and test datasets are a representative example of real-world magnet operation; this makes the good result obtained very promising for future applications in this context.

Keywords: deep neural network, magnetic modelling, measurement and empirical software engineering, NARX

Procedia PDF Downloads 129

Authors: Khyamling Parane, B. M. Prabhu Prasad, Basavaraj Talawar

Abstract:

Fast simulations are critical in reducing time to market in CMPs and SoCs. Several simulators have been used to evaluate the performance and power consumed by Network-on-Chips. Researchers and designers rely upon these simulators for design space exploration of NoC architectures. Our experiments show that simulating large NoC topologies take hours to several days for completion. To speed up the simulations, it is necessary to investigate and optimize the hotspots in simulator source code. Among several simulators available, we choose Booksim2.0, as it is being extensively used in the NoC community. In this paper, we analyze the cache and memory system behaviour of Booksim2.0 to accurately monitor input dependent performance bottlenecks. Our measurements show that cache and memory usage patterns vary widely based on the input parameters given to Booksim2.0. Based on these measurements, the cache configuration having least misses has been identified. To further reduce the cache misses, we use software optimization techniques such as removal of unused functions, loop interchanging and replacing post-increment operator with pre-increment operator for non-primitive data types. The cache misses were reduced by 18.52%, 5.34% and 3.91% by employing above technology respectively. We also employ thread parallelization and vectorization to improve the overall performance of Booksim2.0. The OpenMP programming model and SIMD are used for parallelizing and vectorizing the more time-consuming portions of Booksim2.0. Speedups of 2.93x and 3.97x were observed for the Mesh topology with 30 × 30 network size by employing thread parallelization and vectorization respectively.

Keywords: cache behaviour, network-on-chip, performance profiling, vectorization

Procedia PDF Downloads 196

Authors: Hichem Sassi

Abstract:

Pneumonia, being a respiratory infection, has garnered global attention due to its rapid transmission and relatively high mortality rates. Timely detection and treatment play a crucial role in significantly reducing mortality associated with pneumonia. Presently, X-ray diagnosis stands out as a reasonably effective method. However, the manual scrutiny of a patient's X-ray chest radiograph by a proficient practitioner usually requires 5 to 15 minutes. In situations where cases are concentrated, this places immense pressure on clinicians for timely diagnosis. Relying solely on the visual acumen of imaging doctors proves to be inefficient, particularly given the low speed of manual analysis. Therefore, the integration of artificial intelligence into the clinical image diagnosis of pneumonia becomes imperative. Additionally, AI recognition is notably rapid, with convolutional neural networks (CNNs) demonstrating superior performance compared to human counterparts in image identification tasks. To conduct our study, we utilized a dataset comprising chest X-ray images obtained from Kaggle, encompassing a total of 5216 training images and 624 test images, categorized into two classes: normal and pneumonia. Employing five mainstream network algorithms, we undertook a comprehensive analysis to classify these diseases within the dataset, subsequently comparing the results. The integration of artificial intelligence, particularly through improved network architectures, stands as a transformative step towards more efficient and accurate clinical diagnoses across various medical domains.

Keywords: deep learning, computer vision, pneumonia, models, comparative study

Procedia PDF Downloads 64

Authors: John Gibbs, Benjamin Flanders, Dylan Pozorski, Weixuan Liu

Abstract:

Weight Painting–adjusting the influence a skeletal joint has on a given vertex in a character mesh–is an arduous and time con- suming part of the 3D animation pipeline. This process generally requires a trained technical animator and many hours of work to complete. Our skiNNer plug-in, which works within Autodesk’s Maya 3D animation software, uses Machine Learning and data pro- cessing techniques to create a deep neural network model that can accomplish the weight painting task in seconds rather than hours for bipedal quasi-humanoid character meshes. In order to create a properly trained network, a number of challenges were overcome, including curating an appropriately large data library, managing an arbitrary 3D mesh size, handling arbitrary skeletal architectures, accounting for extreme numeric values (most data points are near 0 or 1 for weight maps), and constructing an appropriate neural network model that can properly capture the high frequency alter- ation between high weight values (near 1.0) and low weight values (near 0.0). The arrived at neural network model is a cross between a traditional CNN, deep residual network, and fully dense network. The resultant network captures the unusually hard-edged features of a weight map matrix, and produces excellent results on many bipedal models.

Keywords: 3d animation, animation, character, rigging, skinning, weight painting, machine learning, artificial intelligence, neural network, deep neural network

Procedia PDF Downloads 270

Authors: Bander Alzahrani

Abstract:

Information-centric networking (ICN) using architectures such as Publish-Subscribe Internet Routing Paradigm (PSIRP) is one of the promising candidates for a future Internet, has recently been under the spotlight by the research community to investigate the possibility of redesigning the current Internet architecture to solve many issues such as routing scalability, security, and quality of services issues.. The Bloom filter-based forwarding is a source-routing approach that is used in the PSIRP architecture. This mechanism is vulnerable to brute force attacks which may lead to denial-of-service (DoS) attacks. In this work, we present a new forwarding approach that keeps the advantages of Bloom filter-based forwarding while mitigates attacks on the forwarding mechanism. In practice, we introduce a special type of forwarding nodes called Edge-FW to be placed at the edge of the network. The role of these node is to add an extra security layer by validating and inspecting packets at the edge of the network against brute-force attacks and check whether the packet contains a legitimate forwarding identifier (FId) or not. We leverage Certificateless Aggregate Signature (CLAS) scheme with a small size of 64-bit which is used to sign the FId. Hence, this signature becomes bound to a specific FId. Therefore, malicious nodes that inject packets with random FIds will be easily detected and dropped at the Edge-FW node when the signature verification fails. Our preliminary security analysis suggests that with the proposed approach, the forwarding plane is able to resist attacks such as DoS with very high probability.

Keywords: bloom filter, certificateless aggregate signature, denial-of-service, information centric network

Procedia PDF Downloads 197

Authors: Nick Weir, Robert Stevens, Alan Hargreaves, Martin McGinnity, Chris Tinsley

Abstract:

Hydrophobic materials have previously demonstrated the ability to elevate cell-cell interactions and promote the formation of neural networks whilst aligned nanofibers demonstrate the ability to induce extensive neurite outgrowth in an aligned manner. Hydrophobic materials typically elicit an immune response upon implantation and thus materials used for implantation are typically hydrophilic. Poly-L-lactic acid (PLLA) is a hydrophobic, non-immunogenic, FDA approved material that can be electrospun to form aligned nanofibers. Primary rat cortical neurons cultured for 10 days on aligned PLLA nanofibers formed 3D cell clusters, approximately 800 microns in diameter. Neurites that extended from these clusters were highly aligned due to the alignment of the nanofibers they were cultured upon and fasciculation was also evident. Plasma treatment of the PLLA nanofibers prior to seeding of cells significantly reduced the hydrophobicity and abolished the cluster formation and neurite fasciculation, whilst reducing the extent and directionality of neurite outgrowth; it is proposed that hydrophobicity induces the changes to cellular behaviors. Aligned PLLA nanofibers induced the formation of a structure that mimics the grey-white matter compartmentalization that is observed in vivo and thus represents a step forward in generating organoids or biomaterial-based implants. Upon implantation into the brain, the biomaterial architectures described here may provide a useful platform for both brain repair and brain remodeling initiatives.

Keywords: hydrophobicity, nanofibers, neurite fasciculation, neurite outgrowth, PLLA

Procedia PDF Downloads 160

Authors: S. Jaramillo, Y. Montoya, W. Agudelo, J. Bustamante

Abstract:

Cardiovascular diseases (CVD) are related to affectations of the heart and blood vessels, within these are pathologies such as coronary or peripheral heart disease, caused by the narrowing of the vessel wall (atherosclerosis), which is related to the accumulation of Low-Density Lipoproteins (LDL) in the arterial walls that leads to a progressive reduction of the lumen of the vessel and alterations in blood perfusion. Currently, the main therapeutic strategy for this type of alteration is drug treatment with statins, which inhibit the enzyme 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMG-CoA reductase), responsible for modulating the rate of cholesterol production and other isoprenoids in the mevalonate pathway. This enzyme induces the expression of LDL receptors in the liver, increasing their number on the surface of liver cells, reducing the plasma concentration of cholesterol. On the other hand, when the blood vessel presents stenosis, a surgical procedure with vascular implants is indicated, which are used to restore circulation in the arterial or venous bed. Among the materials used for the development of vascular implants are Dacron® and Teflon®, which perform the function of re-waterproofing the circulatory circuit, but due to their low biocompatibility, they do not have the ability to promote remodeling and tissue regeneration processes. Based on this, the present research proposes the development of a hydrolyzed collagen and polyethylene oxide electrospun membrane reinforced with medium and high-intensity statins, so that in future research it can favor tissue remodeling processes from its microarchitecture.

Keywords: atherosclerosis, medium and high-intensity statins, microarchitecture, electrospun membrane

Procedia PDF Downloads 134

Authors: Revoti Prasad Bora, Saurabh Yadav, Nikita Katyal

Abstract:

In the context of educational data mining, the use case of extracting information from images containing both text and diagrams is of high importance. Hence, document analysis requires the extraction of diagrams from such images and processes the text and diagrams separately. To the author’s best knowledge, none among plenty of approaches for extracting tables, figures, etc., suffice the need for real-time processing with high accuracy as needed in multiple applications. In the education domain, diagrams can be of varied characteristics viz. line-based i.e. geometric diagrams, chemical bonds, mathematical formulas, etc. There are two broad categories of approaches that try to solve similar problems viz. traditional computer vision based approaches and deep learning approaches. The traditional computer vision based approaches mainly leverage connected components and distance transform based processing and hence perform well in very limited scenarios. The existing deep learning approaches either leverage YOLO or faster-RCNN architectures. These approaches suffer from a performance-accuracy tradeoff. This paper proposes a U-Net based architecture that formulates the diagram extraction as a segmentation problem. The proposed method provides similar accuracy with a much faster extraction time as compared to the mentioned state-of-the-art approaches. Further, the segmentation mask in this approach allows the extraction of diagrams of irregular shapes.

Keywords: computer vision, deep-learning, educational data mining, faster-RCNN, figure extraction, image segmentation, real-time document analysis, text extraction, U-Net, YOLO

Procedia PDF Downloads 137