Search results for: multimodal fusion classifier

Authors: Yulan Wu

Abstract:

The spread of fake news on social media has posed significant societal harm to the public and the nation, with its threats spanning various domains, including politics, economics, health, and more. News on social media often covers multiple domains, and existing models studied by researchers and relevant organizations often perform well on datasets from a single domain. However, when these methods are applied to social platforms with news spanning multiple domains, their performance significantly deteriorates. Existing research has attempted to enhance the detection performance of multi-domain datasets by adding single-domain labels to the data. However, these methods overlook the fact that a news article typically belongs to multiple domains, leading to the loss of domain knowledge information contained within the news text. To address this issue, research has found that news records in different domains often use different vocabularies to describe their content. In this paper, we propose a fake news detection framework that combines domain knowledge and expert knowledge. Firstly, it utilizes an unsupervised domain discovery module to generate a low-dimensional vector for each news article, representing domain embeddings, which can retain multi-domain knowledge of the news content. Then, a feature extraction module uses the domain embeddings discovered through unsupervised domain knowledge to guide multiple experts in extracting news knowledge for the total feature representation. Finally, a classifier is used to determine whether the news is fake or not. Experiments show that this approach can improve multi-domain fake news detection performance while reducing the cost of manually labeling domain labels.

Keywords: fake news, deep learning, natural language processing, multiple domains

Procedia PDF Downloads 42

Authors: Massimo Miccoli, Luca Marangoni, Alberto Aniello Scaringi, Alessandro Marceddu, Alessandro Amicone

Abstract:

The susceptibility of deep neural networks (DNNs) to adversarial manipulations is a recognized challenge within the computer vision domain. Adversarial examples, crafted by adding subtle yet malicious alterations to benign images, exploit this vulnerability. Various defense strategies have been proposed to safeguard DNNs against such attacks, stemming from diverse research hypotheses. Building upon prior work, our approach involves the utilization of autoencoder models. Autoencoders, a type of neural network, are trained to learn representations of training data and reconstruct inputs from these representations, typically minimizing reconstruction errors like mean squared error (MSE). Our autoencoder was trained on a dataset of benign examples; learning features specific to them. Consequently, when presented with significantly perturbed adversarial examples, the autoencoder exhibited high reconstruction errors. The architecture of the autoencoder was tailored to the dimensions of the images under evaluation. We considered various image sizes, constructing models differently for 256x256 and 512x512 images. Moreover, the choice of the computer vision model is crucial, as most adversarial attacks are designed with specific AI structures in mind. To mitigate this, we proposed a method to replace image-specific dimensions with a structure independent of both dimensions and neural network models, thereby enhancing robustness. Our multi-modal autoencoder reconstructs the spectral representation of images across the red-green-blue (RGB) color channels. To validate our approach, we conducted experiments using diverse datasets and subjected them to adversarial attacks using models such as ResNet50 and ViT_L_16 from the torch vision library. The autoencoder extracted features used in a classification model, resulting in an MSE (RGB) of 0.014, a classification accuracy of 97.33%, and a precision of 99%.

Keywords: adversarial attacks, malicious images detector, binary classifier, multimodal transformer autoencoder

Procedia PDF Downloads 48

Authors: Thanh Noi Phan, Martin Kappas, Jan Degener

Abstract:

The currently launched Sentinel 2 (S2) satellite (June, 2015) bring a great potential and opportunities for land use/cover map applications, due to its fine spatial resolution multispectral as well as high temporal resolutions. So far, there are handful studies using S2 real data for land cover classification. Especially in northern Vietnam, to our best knowledge, there exist no studies using S2 data for land cover map application. The aim of this study is to provide the preliminary result of land cover classification using Sentinel -2 data with a rising state – of – art classifier, Random Forest. A case study with heterogeneous land use/cover in the eastern of Hanoi Capital – Vietnam was chosen for this study. All 10 spectral bands of 10 and 20 m pixel size of S2 images were used, the 10 m bands were resampled to 20 m. Among several classified algorithms, supervised Random Forest classifier (RF) was applied because it was reported as one of the most accuracy methods of satellite image classification. The results showed that the red-edge and shortwave infrared (SWIR) bands play an important role in land cover classified results. A very high overall accuracy above 90% of classification results was achieved.

Keywords: classify algorithm, classification, land cover, random forest, sentinel 2, Vietnam

Procedia PDF Downloads 353

Authors: Pengfei Meng, Shuangcheng Jia, Qian Li

Abstract:

We proposed a real-time high-precision semantic segmentation network based on a multi-resolution feature fusion module, the auxiliary feature extracting module, upsampling module, and atrous spatial pyramid pooling (ASPP) module. We designed a feature fusion structure, which is integrated with sufficient features of different resolutions. We also studied the effect of side-branch structure on the network and made discoveries. Based on the discoveries about the side-branch of the network structure, we used a side-branch auxiliary feature extraction layer in the network to improve the effectiveness of the network. We also designed upsampling module, which has better results than the original upsampling module. In addition, we also re-considered the locations and number of atrous spatial pyramid pooling (ASPP) modules and modified the network structure according to the experimental results to further improve the effectiveness of the network. The network presented in this paper takes the backbone network of Bisenetv2 as a basic network, based on which we constructed a network structure on which we made improvements. We named this network deep multiple-resolution bilateral networks for real-time, referred to as DMBR-Net. After experimental testing, our proposed DMBR-Net network achieved 81.2% mIoU at 119FPS on the Cityscapes validation dataset, 80.7% mIoU at 109FPS on the CamVid test dataset, 29.9% mIoU at 78FPS on the COCOStuff test dataset. Compared with all lightweight real-time semantic segmentation networks, our network achieves the highest accuracy at an appropriate speed.

Keywords: multi-resolution feature fusion, atrous convolutional, bilateral networks, pyramid pooling

Procedia PDF Downloads 112

Authors: Chusak Thanawattano, Roongroj Bhidayasiri

Abstract:

This paper proposes a novel feature set utilized for classifying the Parkinson tremor and essential tremor. Ten ET and ten PD subjects are asked to perform kinetic, postural and resting tests. The empirical mode decomposition (EMD) is used to decompose collected tremor signal to a set of intrinsic mode functions (IMF). The IMFs are used for reconstructing representative signals. The feature set is composed of peak frequencies of IMFs and reconstructed signals. Hypothesize that the dominant frequency components of subjects with PD and ET change in different directions for different tests, difference of peak frequencies of IMFs and reconstructed signals of pairwise based tests (kinetic-resting, kinetic-postural and postural-resting) are considered as potential features. Sets of features are used to train and test by classifier including the quadratic discriminant classifier (QLC) and the support vector machine (SVM). The best accuracy, the best sensitivity and the best specificity are 90%, 87.5%, and 92.86%, respectively.

Keywords: tremor, Parkinson, essential tremor, empirical mode decomposition, quadratic discriminant, support vector machine, peak frequency, auto-regressive, spectrum estimation

Procedia PDF Downloads 415

Authors: Sehreen Moorat, Mussarat Lakho

Abstract:

A method of artificial intelligence using digital mammograms data has been proposed in this paper for detection of breast cancer. Many researchers have developed techniques for the early detection of breast cancer; the early diagnosis helps to save many lives. The detection of breast cancer through mammography is effective method which detects the cancer before it is felt and increases the survival rate. In this paper, we have purposed image processing technique for enhancing the image to detect the graphical table data and markings. Texture features based on Gray-Level Co-Occurrence Matrix and intensity based features are extracted from the selected region. For classification purpose, neural network based supervised classifier system has been used which can discriminate between benign and malignant. Hence, 68 digital mammograms have been used to train the classifier. The obtained result proved that automated detection of breast cancer is beneficial for early diagnosis and increases the survival rates of breast cancer patients. The proposed system will help radiologist in the better interpretation of breast cancer.

Keywords: medical imaging, cancer, processing, neural network

Procedia PDF Downloads 236

Authors: Raymond Xu, Ashley Hua, Andrew Wang, Yuru Lin

Abstract:

During COVID-19, the depression rate has increased dramatically. Young adults are most vulnerable to the mental health effects of the pandemic. Lower-income families have a higher ratio to be diagnosed with depression than the general population, but less access to clinics. This research aims to achieve early depression detection at low cost, large scale, and high accuracy with an interdisciplinary approach by incorporating clinical practices defined by American Psychiatric Association (APA) as well as multimodal AI framework. The proposed approach detected the nine depression symptoms with Natural Language Processing sentiment analysis and a symptom-based Lexicon uniquely designed for young adults. The experiments were conducted on the multimedia survey results from adolescents and young adults and unbiased Twitter communications. The result was further aggregated with the facial emotional cues analyzed by the Convolutional Neural Network on the multimedia survey videos. Five experiments each conducted on 10k data entries reached consistent results with an average accuracy of 88.31%, higher than the existing natural language analysis models. This approach can reach 300+ million daily active Twitter users and is highly accessible by low-income populations to promote early depression detection to raise awareness in adolescents and young adults and reveal complementary cues to assist clinical depression diagnosis.

Keywords: artificial intelligence, COVID-19, depression detection, psychiatric disorder

Procedia PDF Downloads 108

Authors: Mirjana Ruppel, Rajendra Persad, Amit Bahl, Sanja Dogramadzi, Chris Melhuish, Lyndon Smith

Abstract:

Multimodal image registration is a profoundly complex task which is why deep learning has been used widely to address it in recent years. However, two main challenges remain: Firstly, the lack of ground truth data calls for an unsupervised learning approach, which leads to the second challenge of defining a feasible loss function that can compare two images of different modalities to judge their level of alignment. To avoid this issue altogether we implement a generative adversarial network consisting of two registration networks GAB, GBA and two discrimination networks DA, DB connected by spatial transformation layers. GAB learns to generate a deformation field which registers an image of the modality B to an image of the modality A. To do that, it uses the feedback of the discriminator DB which is learning to judge the quality of alignment of the registered image B. GBA and DA learn a mapping from modality A to modality B. Additionally, a cycle-consistency loss is implemented. For this, both registration networks are employed twice, therefore resulting in images ˆA, ˆB which were registered to ˜B, ˜A which were registered to the initial image pair A, B. Thus the resulting and initial images of the same modality can be easily compared. A dataset of liver CT and MRI was used to evaluate the quality of our approach and to compare it against learning and non-learning based registration algorithms. Our approach leads to dice scores of up to 0.80 ± 0.01 and is therefore comparable to and slightly more successful than algorithms like SimpleElastix and VoxelMorph.

Keywords: cycle consistency, deformable multimodal image registration, deep learning, GAN

Procedia PDF Downloads 104

Authors: Ebipatei Victoria Tunyan, T. A. Cao, Cheol Young Ock

Abstract:

Detecting subjectively biased statements is a vital task. This is because this kind of bias, when present in the text or other forms of information dissemination media such as news, social media, scientific texts, and encyclopedias, can weaken trust in the information and stir conflicts amongst consumers. Subjective bias detection is also critical for many Natural Language Processing (NLP) tasks like sentiment analysis, opinion identification, and bias neutralization. Having a system that can adequately detect subjectivity in text will boost research in the above-mentioned areas significantly. It can also come in handy for platforms like Wikipedia, where the use of neutral language is of importance. The goal of this work is to identify the subjectively biased language in text on a sentence level. With machine learning, we can solve complex AI problems, making it a good fit for the problem of subjective bias detection. A key step in this approach is to train a classifier based on BERT (Bidirectional Encoder Representations from Transformers) as upstream model. BERT by itself can be used as a classifier; however, in this study, we use BERT as data preprocessor as well as an embedding generator for a Bi-LSTM (Bidirectional Long Short-Term Memory) network incorporated with attention mechanism. This approach produces a deeper and better classifier. We evaluate the effectiveness of our model using the Wiki Neutrality Corpus (WNC), which was compiled from Wikipedia edits that removed various biased instances from sentences as a benchmark dataset, with which we also compare our model to existing approaches. Experimental analysis indicates an improved performance, as our model achieved state-of-the-art accuracy in detecting subjective bias. This study focuses on the English language, but the model can be fine-tuned to accommodate other languages.

Keywords: subjective bias detection, machine learning, BERT–BiLSTM–Attention, text classification, natural language processing

Procedia PDF Downloads 101

Authors: Flaurence Benjamain, Michel Casperance

Abstract:

We present in this paper, first, a comparative study of three mathematical theories to achieve the fusion of information sources. This study aims to identify the characteristics inherent in theories of possibilities, belief functions (DST) and plausible and paradoxical reasoning to establish a strategy of choice that allows us to adopt the most appropriate theory to solve a problem of fusion in order, taking into account the acquired information and imperfections that accompany them. Using the new theory of plausible and paradoxical reasoning, also called Dezert-Smarandache Theory (DSmT), to fuse information multi-sources needs, at first step, the generation of the composites events witch is, in general, difficult. Thus, we present in this paper a new approach to construct pertinent paradoxical classes based on gray levels histograms, which also allows to reduce the cardinality of the hyper-powerset. Secondly, we developed a new technique for order and coding generalized focal elements. This method is exploited, in particular, to calculate the cardinality of Dezert and Smarandache. Then, we give an experimentation of classification of a remote sensing image that illustrates the given methods and we compared the result obtained by the DSmT with that resulting from the use of the DST and theory of possibilities.

Keywords: segmentation, image, approach, vision computing

Procedia PDF Downloads 253

Authors: Daniel Schimmelpfennig

Abstract:

This paper attempts to delineate the idea to curate the leveling up for whole-systems change. Curation is the act fo select, organize, look after, or present information from a professional point of view through expert knowledge. The trans-paradigmatic, trans-contextual, trans-disciplinary, trans-perspective of trans-media futures studies hopes to enable a move from a monochrome intellectual pursuit towards breathing a higher dimensionality. Progressing to the next level to equip actors for whole-systems change is in consideration of the commonly known symptoms of our time as well as in anticipation of future challenges, both a necessity and desirability. Systems of collective intelligence could potentially scale regenerative, adaptive, and anticipatory capacities. How could such a curation then be enacted and implemented, to initiate the process of leveling-up? The suggestion here is to focus on the metasystem transition, the bio-digital fusion, namely, by merging neurosciences, the ontological design of money as our operating system, and our understanding of the billions of years of time-proven permutations in nature, biomimicry, and biological metaphors like symbiogenesis. Evolutionary cybernetics accompanies the process of whole-systems change.

Keywords: bio-digital fusion, evolutionary cybernetics, metasystem transition, symbiogenesis, transmedia futures studies

Procedia PDF Downloads 122

Authors: A. Alem, Y. Dahmani, A. Hadjali, A. Boualem

Abstract:

Managing the problem of the conflict, either by using the Dempster-Shafer theory, or by the application of the fusion process to push researchers in recent years to find ways to get to make best decisions especially; for information systems, vision, robotic and wireless sensor networks. In this paper we are interested to take account of the conflict in the combination step that took the conflict into account and tries to manage such a way that it does not influence the decision step, the conflict what from reliable sources. According to [1], the conflict lead to erroneous decisions in cases where was with strong degrees between sources of information, if the conflict is more than the maximum of the functions of belief mass K > max1...n (mi (A)), then the decision becomes impossible. We will demonstrate in this paper that the multiplication of mass functions by coefficients of reliability is a decreasing function; it leads to the reduction of conflict and a good decision. The definition of reliability coefficients accurately and multiply them by the mass functions of each information source to resolve the conflict and allow deciding whether the degree of conflict. The evaluation of this technique is done by a use case; a comparison of the combination of springs with a maximum conflict without, and with reliability coefficients.

Keywords: Dempster-Shafer theory, fusion process, conflict managing, reliability factors, decision

Procedia PDF Downloads 396

Authors: Orsolya Horvath, Laszlo Deres, Krisztian Eros, Katalin Ordog, Tamas Habon, Balazs Sumegi, Kalman Toth, Robert Halmosi

Abstract:

Introduction: Oxidative stress induces an imbalance in mitochondrial fusion and fission processes, finally leading to cell death. The two antioxidant molecules, BGP-15 and L2286 have beneficial effects on mitochondrial functions and on cellular oxidative stress response. In this work, we studied the effects of these compounds on the processes of mitochondrial quality control. Methods: We used H9c2 cardiomyoblast and isolated neonatal rat cardiomyocytes (NRCM) for the experiments. The concentration of stressors and antioxidants was beforehand determined with MTT test. We applied 1-Methyl-3-nitro-1-nitrosoguanidine (MNNG) in 125 µM, 400 µM and 800 µM concentrations for 4 and 8 hours on H9c2 cells. H₂O₂ was applied in 150 µM and 300 µM concentration for 0.5 and 4 hours on both models. L2286 was administered in 10 µM, while BGP-15 in 50 µM doses. Cellular levels of the key proteins playing role in mitochondrial dynamics were measured in Western blot samples. For the analysis of mitochondrial network dynamics, we applied electron microscopy and immunocytochemistry. Results: Due to MNNG treatment the level of fusion proteins (OPA1, MFN2) decreased, while the level of fission protein DRP1 elevated markedly. The levels of fusion proteins OPA1 and MNF2 increased in the L2286 and BGP-15 treated groups. During the 8 hour treatment period, the level of DRP1 also increased in the treated cells (p < 0.05). In the H₂O₂ stressed cells, administration of L2286 increased the level of OPA1 in both H9c2 and NRCM models. MFN2 levels in isolated neonatal rat cardiomyocytes raised considerably due to BGP-15 treatment (p < 0.05). L2286 administration decreased the DRP1 level in H9c2 cells (p < 0.05). We observed that the H₂O₂-induced mitochondrial fragmentation could be decreased by L2286 treatment. Conclusion: Our results indicated that the PARP-inhibitor L2286 has beneficial effect on mitochondrial dynamics during oxidative stress scenario, and also in the case of directly induced DNA damage. We could make the similar conclusions in case of BGP-15 administration, which, via reducing ROS accumulation, propagates fusion processes, this way aids preserving cellular viability. Funding: GINOP-2.3.2-15-2016-00049; GINOP-2.3.2-15-2016-00048; GINOP-2.3.3-15-2016-00025; EFOP-3.6.1-16-2016-00004; ÚNKP-17-4-I-PTE-209

Keywords: H9c2, mitochondrial dynamics, neonatal rat cardiomyocytes, oxidative stress

Procedia PDF Downloads 125

Authors: Fatemeh Mirakhorli

Abstract:

Porosity is the main issue during welding of aluminum alloys, and surface cleaning has a critical influence to reduce the porosity level by removing the oxidized surface layer before fusion welding. Developing an optimum and economical surface cleaning method has an enormous benefit for aluminum welding industries to reduce costs related to repairing and repeating welds as well as increasing the mechanical properties of the joints. In this study, several mechanical and chemical surface cleaning methods were examined for butt joint welding of 2 mm thick AA6xxx alloys using ER5556 filler metal. The effects of each method on porosity formation and tensile properties are evaluated. It has been found that, compared to the conventional mechanical cleaning method, the use of chemical cleaning leads to an important reduction in porosity level even after a significant delay between cleaning and welding. The effect of the higher porosity level in the fusion zone to reduce the tensile strength of the welds is shown.

Keywords: gas metal arc welding (GMAW), aluminum alloy, surface cleaning, porosity formation, mechanical property

Procedia PDF Downloads 112

Authors: Han Xue, Zhang Lanyue

Abstract:

In order to seek more effective feature extraction technology, feature extraction method based on MFCC combined with vector hydrophone is exposed in the paper. The sound pressure signal and particle velocity signal of two kinds of ships are extracted by using MFCC and its evolution form, and the extracted features are fused by using fisher-ratio and correlated distance criterion. The features are then identified by BP neural network. The results showed that MFCC, First-Order Differential MFCC and Second-Order Differential MFCC features can be used as effective features for recognition of underwater targets, and the fusion feature can improve the recognition rate. Moreover, the results also showed that the recognition rate of the particle velocity signal is higher than that of the sound pressure signal, and it reflects the superiority of vector signal processing.

Keywords: vector information, MFCC, differential MFCC, fusion feature, BP neural network

Procedia PDF Downloads 494

Authors: Fedili Benamar, Beloulou Mohamed Lamine, Ouahes Hassane, Ghattas Samir

Abstract:

 Background and aims: Peripheral regional anesthesia has been integrated into most analgesia protocols for total knee arthroplasty which considered among the most painful surgeries with a huge potential for chronicization. The adductor canal block (ACB) has gained popularity. Similarly, the IPACK block has been described to provide analgesia of the posterior knee capsule. This study aimed to evaluate the analgesic efficacy of this block in patients undergoing primary PTG. Methods: 90 patients were randomized to receive either an IPACK, an anterior sciatic block, or a sham block (30 patients in each group + multimodal analgesia and a catheter in the KCA adductor canal). GROUP 1 KCA GROUP 2 KCA+BSA GROUP 3 KCA+IPACK The analgesic blocks were done under echo-guidance preoperatively respecting the safety rules, the dose administered was 20 cc of ropivacaine 0.25% was used. We were to assess posterior knee pain 6 hours after surgery. Other endpoints included quality of recovery after surgery, pain scores, opioid requirements (PCA morphine)(EPI info 7.2 analysis). Results: -groups were matched -A predominance of women (4F/1H). -average age: 68 +/-7 years -the average BMI =31.75 kg/m2 +/- 4. -70% of patients ASA2 ,20% ASA3. -The average duration of the intervention: 89 +/- 19 minutes. -Morphine consumption (PCA) significantly higher in group 1 (16mg) & group 2 (8mg) group 3 (4mg) - The groups were matched . -There was a correlation between the use of the ipack block and postoperative pain Conclusions :In a multimodal analgesic protocol, the addition of IPACK block decreased pain scores and morphine consumption ,

Keywords: regional anesthesia, analgesia, total knee arthroplasty, the adductor canal block (acb), the ipack block, pain

Procedia PDF Downloads 46

Authors: Marina Charalambous, Phivos Phylactou, Thekla Elriz, Loukia Psychogios, Jean-Marie Annoni

Abstract:

Background: Evidence-based practices for the evaluation and treatment of people with aphasia (PWA) in Greek are mainly impairment-based. Functional and multimodal communication is usually under assessed and neglected by clinicians. This study explores the adaptation and psychometric testing of the Greek (GR) version of The Scenario Test. The Scenario Test assesses the everyday functional communication of PWA in an interactive multimodal communication setting with the support of an active communication facilitator. Aims: To define the reliability and validity of The Scenario Test GR and discuss its clinical value. Methods & Procedures: The Scenario Test-GR was administered to 54 people with chronic stroke (6+ months post-stroke): 32 PWA and 22 people with stroke without aphasia. Participants were recruited from Greece and Cyprus. All measures were performed in an interview format. Standard psychometric criteria were applied to evaluate reliability (internal consistency, test-retest, and interrater reliability) and validity (construct and known – groups validity) of the Scenario Test GR. Video analysis was performed for the qualitative examination of the communication modes used. Outcomes & Results: The Scenario Test-GR shows high levels of reliability and validity. High scores of internal consistency (Cronbach’s α = .95), test-retest reliability (ICC = .99), and interrater reliability (ICC = .99) were found. Interrater agreement in scores on individual items fell between good and excellent levels of agreement. Correlations with a tool measuring language function in aphasia (the Aphasia Severity Rating Scale of the Boston Diagnostic Aphasia Examination), a measure of functional communication (the Communicative Effectiveness Index), and two instruments examining the psychosocial impact of aphasia (the Stroke and Aphasia Quality of Life questionnaire and the Aphasia Impact Questionnaire) revealed good convergent validity (all ps< .05). Results showed good known – groups validity (Mann-Whitney U = 96.5, p < .001), with significantly higher scores for participants without aphasia compared to those with aphasia. Conclusions: The psychometric qualities of The Scenario Test-GR support the reliability and validity of the tool for the assessment of functional communication for Greek-speaking PWA. The Scenario Test-GR can be used to assess multimodal functional communication, orient aphasia rehabilitation goal setting towards the activity and participation level, and be used as an outcome measure of everyday communication. Future studies will focus on the measurement of sensitivity to change in PWA with severe non-fluent aphasia.

Keywords: the scenario test GR, functional communication assessment, people with aphasia (PWA), tool validation

Procedia PDF Downloads 106

Authors: Hadri Hussain, M.Nasir Ibrahim, Chee-Ming Ting, Mariani Idroas, Fuad Numan, Alias Mohd Noor

Abstract:

A biometric security system has become an important application in client identification and verification system. A conventional biometric system is normally based on unimodal biometric that depends on either behavioural or physiological information for authentication purposes. The behavioural biometric depends on human body biometric signal (such as speech) and biosignal biometric (such as electrocardiogram (ECG) and phonocardiogram or heart sound (HS)). The speech signal is commonly used in a recognition system in biometric, while the ECG and the HS have been used to identify a person’s diseases uniquely related to its cluster. However, the conventional biometric system is liable to spoof attack that will affect the performance of the system. Therefore, a multimodal biometric security system is developed, which is based on biometric signal of ECG, HS, and speech. The biosignal data involved in the biometric system is initially segmented, with each segment Mel Frequency Cepstral Coefficients (MFCC) method is exploited for extracting the feature. The Hidden Markov Model (HMM) is used to model the client and to classify the unknown input with respect to the modal. The recognition system involved training and testing session that is known as client identification (CID). In this project, twenty clients are tested with the developed system. The best overall performance at 44 kHz was 93.92% for ECG and the worst overall performance was ECG at 88.47%. The results were compared to the best overall performance at 44 kHz for (20clients) to increment of clients, which was 90.00% for HS and the worst overall performance falls at ECG at 79.91%. It can be concluded that the difference multimodal biometric has a substantial effect on performance of the biometric system and with the increment of data, even with higher frequency sampling, the performance still decreased slightly as predicted.

Keywords: electrocardiogram, phonocardiogram, hidden markov model, mel frequency cepstral coeffiecients, client identification

Procedia PDF Downloads 258

Authors: Khaled Nagaty, Heba Nagaty, Gerard McKee

Abstract:

A signature is a behavioral biometric that is used for authenticating users in most financial and legal transactions. Signatures can be easily forged by skilled forgers. Therefore, it is essential to verify whether a signature is genuine or forged. The aim of any signature verification algorithm is to accommodate the differences between signatures of the same person and increase the ability to discriminate between signatures of different persons. This work presented in this paper proposes an automatic signature verification system to indicate whether a signature is genuine or not. The system comprises four phases: (1) The pre-processing phase in which image scaling, binarization, image rotation, dilation, thinning, and connecting ridge breaks are applied. (2) The feature extraction phase in which global and local features are extracted. The local features are minutiae points, curvature orientation, and curve plateau. The global features are signature area, signature aspect ratio, and Hu moments. (3) The post-processing phase, in which false minutiae are removed. (4) The classification phase in which features are enhanced before feeding it into the classifier. k-nearest neighbors and support vector machines are used. The classifier was trained on a benchmark dataset to compare the performance of the proposed offline signature verification system against the state-of-the-art. The accuracy of the proposed system is 92.3%.

Keywords: signature, ridge breaks, minutiae, orientation

Procedia PDF Downloads 124

Authors: Saulutė Juzelėnienė, Skirmantė Šarkauskienė

Abstract:

In this paper, it is aimed to analyse the multimodal representations of the concepts of LOVE and PASSION in Lithuanian graphic novel “Gertrūda”, by Gerda Jord. The research is based on the earlier findings by Forceville (2005), Eerden (2009) as well as insights made by Shihara and Matsunaka (2009) and Kövecses (2000). The domains of target and source of LOVE and PASSION metaphors in comics are expressed by verbal and non-verbal cues. The analysis of non-verbal cues adopts the concepts of rune and indexes. A pictorial rune is a graphic representation of an object that does not exist in reality in comics, such as lines, dashes, text "balloons", and pictorial index – a graphically represented object of reality, a real symptom expressing a certain emotion, such as a wide smile, furrowed eyebrows, etc. Indexes are often hyperbolized in comics. The research revealed that most frequent source domains are CLOSINESS/UNITY, NATURAL/ PHYSICAL FORCE, VALUABLE OBJECT, PRESSURE. The target is the emotion of LOVE/PASSION which belongs to a more abstract domain of psychological experience. In this kind of metaphor, the picture can be interpreted as representing the emotion of happiness. Data are taken from Lithuanian comic books and Internet sites, where comics have been presented. The data and the analysis we are providing in this article aims to reveal that there are pictorial metaphors that manifest conceptual metaphors that are also expressed verbally and that methodological framework constructed for the analysis in the papers by Forceville at all is applicable to other emotions and culture specific pictorial manifestations.

Keywords: multimodal metaphor, conceptual metaphor, comics, graphic novel, concept of love/passion

Procedia PDF Downloads 49

Authors: Yury S. Shpanskiy, Boris V. Kuteev

Abstract:

Development of a fusion-fission hybrid facility based on superconducting conventional tokamak DEMO-FNS runs in Russia since 2013. The main design goal is to reach the technical feasibility and outline prospects of industrial hybrid technologies providing the production of neutrons, fuel nuclides, tritium, high-temperature heat, electricity and subcritical transmutation in Fusion-Fission Hybrid Systems. The facility should operate in a steady-state mode at the fusion power of 40 MW and fission reactions of 400 MW. Major tokamak parameters are the following: major radius R=3.2 m, minor radius a=1.0 m, elongation 2.1, triangularity 0.5. The design provides the neutron wall loading of ~0.2 MW/m², the lifetime neutron fluence of ~2 MWa/m², with the surface area of the active cores and tritium breeding blanket ~100 m². Core plasma modelling showed that the neutron yield ~10¹⁹ n/s is maximal if the tritium/deuterium density ratio is 1.5-2.3. The design of the electromagnetic system (EMS) defined its basic parameters, accounting for the coils strength and stability, and identified the most problematic nodes in the toroidal field coils and the central solenoid. The EMS generates toroidal, poloidal and correcting magnetic fields necessary for the plasma shaping and confinement inside the vacuum vessel. EMC consists of eighteen superconducting toroidal field coils, eight poloidal field coils, five sections of a central solenoid, correction coils, in-vessel coils for vertical plasma control. Supporting structures, the thermal shield, and the cryostat maintain its operation. EMS operates with the pulse duration of up to 5000 hours at the plasma current up to 5 MA. The vacuum vessel (VV) is an all-welded two-layer toroidal shell placed inside the EMS. The free space between the vessel shells is filled with water and boron steel plates, which form the neutron protection of the EMS. The VV-volume is 265 m³, its mass with manifolds is 1800 tons. The nuclear blanket of DEMO-FNS facility was designed to provide functions of minor actinides transmutation, tritium production and enrichment of spent nuclear fuel. The vertical overloading of the subcritical active cores with MA was chosen as prospective. Analysis of the device neutronics and the hybrid blanket thermal-hydraulic characteristics has been performed for the system with functions covering transmutation of minor actinides, production of tritium and enrichment of spent nuclear fuel. A study of FNS facilities role in the Russian closed nuclear fuel cycle was performed. It showed that during ~100 years of operation three FNS facilities with fission power of 3 GW controlled by fusion neutron source with power of 40 MW can burn 98 tons of minor actinides and 198 tons of Pu-239 can be produced for startup loading of 20 fast reactors. Instead of Pu-239, up to 25 kg of tritium per year may be produced for startup of fusion reactors using blocks with lithium orthosilicate instead of fissile breeder blankets.

Keywords: fusion-fission hybrid system, conventional tokamak, superconducting electromagnetic system, two-layer vacuum vessel, subcritical active cores, nuclear fuel cycle

Procedia PDF Downloads 128

Authors: Vimala Rani P, Narasimha Malikarjunan, Mercy Shalinie S

Abstract:

Data Networking was brought forth as an instantiation of information-centric networking. The attackers can send a colossal number of spoofs to take hold of the Pending Interest Table (PIT) named an Interest Flooding attack (IFA) since the in- interests are recorded in the PITs of the intermediate routers until they receive corresponding Data Packets are go beyond the time limit. These attacks can be detrimental to network performance. PIT expiration rate or the Interest satisfaction rate, which cannot differentiate the IFA from attacks, is the criterion Traditional IFA detection techniques are concerned with. Threshold values can casually affect Threshold-based traditional methods. This article proposes an accurate IFA detection mechanism based on a Multiple Feature-based Extreme Learning Machine (MF-ELM). Accuracy of the attack detection can be increased by presenting the entropy of Internet names, Interest satisfaction rate and PIT usage as features extracted in the MF-ELM classifier. Furthermore, we deploy a queue-based hostile Interest prefix mitigation mechanism. The inference of this real-time test bed is that the mechanism can help the network to resist IFA with higher accuracy and efficiency.

Keywords: information-centric network, pending interest table, interest flooding attack, MF-ELM classifier, queue-based mitigation strategy

Procedia PDF Downloads 183

Authors: Yaolin Tian, Shanxiong Chen, Fujia Zhao, Xiaoyu Lin, Hailing Xiong

Abstract:

Ancient books are significant culture inheritors and their background textures convey the potential history information. However, multi-style texture recovery of ancient books has received little attention. Restricted by insufficient ancient textures and complex handling process, the generation of ancient textures confronts with new challenges. For instance, training without sufficient data usually brings about overfitting or mode collapse, so some of the outputs are prone to be fake. Recently, image generation and style transfer based on deep learning are widely applied in computer vision. Breakthroughs within the field make it possible to conduct research upon multi-style texture recovery of ancient books. Under the circumstances, we proposed a network of layout analysis and image fusion system. Firstly, we trained models by using Deep Convolution Generative against Networks (DCGAN) to synthesize multi-style ancient textures; then, we analyzed layouts based on the Position Rearrangement (PR) algorithm that we proposed to adjust the layout structure of foreground content; at last, we realized our goal by fusing rearranged foreground texts and generated background. In experiments, diversified samples such as ancient Yi, Jurchen, Seal were selected as our training sets. Then, the performances of different fine-turning models were gradually improved by adjusting DCGAN model in parameters as well as structures. In order to evaluate the results scientifically, cross entropy loss function and Fréchet Inception Distance (FID) are selected to be our assessment criteria. Eventually, we got model M8 with lowest FID score. Compared with DCGAN model proposed by Radford at el., the FID score of M8 improved by 19.26%, enhancing the quality of the synthetic images profoundly.

Keywords: deep learning, image fusion, image generation, layout analysis

Procedia PDF Downloads 120

Authors: T. Praveenkumar, Kulpreet Singh, Divy Bhanpuriya, M. Saimurugan

Abstract:

This study analysed the classification accuracy for gearbox faults using Machine Learning Techniques. Gearboxes are widely used for mechanical power transmission in rotating machines. Its rotating components such as bearings, gears, and shafts tend to wear due to prolonged usage, causing fluctuating vibrations. Increasing the dependability of mechanical components like a gearbox is hampered by their sealed design, which makes visual inspection difficult. One way of detecting impending failure is to detect a change in the vibration signature. The current study proposes various machine learning algorithms, with aid of these vibration signals for obtaining the fault classification accuracy of an automotive 4-Speed synchromesh gearbox. Experimental data in the form of vibration signals were acquired from a 4-Speed synchromesh gearbox using Data Acquisition System (DAQs). Statistical features were extracted from the acquired vibration signal under various operating conditions. Then the extracted features were given as input to the algorithms for fault classification. Supervised Machine Learning algorithms such as Support Vector Machines (SVM) and unsupervised algorithms such as Deep Feed Forward Neural Network (DFFNN), Deep Belief Networks (DBN) algorithms are used for fault classification. The fusion of DBN & DFFNN classifiers were architected to further enhance the classification accuracy and to reduce the computational complexity. The fault classification accuracy for each algorithm was thoroughly studied, tabulated, and graphically analysed for fused and individual algorithms. In conclusion, the fusion of DBN and DFFNN algorithm yielded the better classification accuracy and was selected for fault detection due to its faster computational processing and greater efficiency.

Keywords: deep belief networks, DBN, deep feed forward neural network, DFFNN, fault diagnosis, fusion of algorithm, vibration signal

Procedia PDF Downloads 87

Authors: Tunde Ope-Davies

Abstract:

Transmediality and platformization as online content-sharing protocols have continued to accentuate the growing impact of the unprecedented digital revolution across the world. The rapid transformation across all sectors as a result of this revolution has continued to spotlight the increasing importance of new media technologies in redefining and reshaping the rhythm and dynamics of our private and public discursive practices. Equally, social and political activities are being impacted daily through the creation and transmission of political discourse content through multi-channel platforms such as mobile telephone communication, social media networks and the internet. It has been observed that digital platforms have become central to the production, processing, and distribution of multimodal social data and cultural content. The platformization paradigm thus underpins our understanding of how digital platforms enhance the production and heterogenous distribution of media and cultural content through these platforms and how this process facilitates socioeconomic and political activities. The use of multiple digital platforms to share and transmit political discourse material synchronously and asynchronously has gained some exciting momentum in the last few years. Nigeria’s 2023 general elections amplified the usage of social media and other online platforms as tools for electioneering campaigns, socio-political mobilizations and civic engagement. The study, therefore, focuses on transmedia and platformed political discourse as a new strategy to promote political candidates and their manifesto in order to mobilize support and woo voters. This innovative transmedia digital discourse model involves a constellation of online texts and images transmitted through different online platforms almost simultaneously. The data for the study was extracted from the 2023 general elections campaigns in Nigeria between January- March 2023 through media monitoring, manual download and the use of software to harvest the online electioneering campaign material. I adopted a discursive-analytic qualitative technique with toolkits drawn from a computer-mediated multimodal discourse paradigm. The study maps the progressive development of digital political discourse in this young democracy. The findings also demonstrate the inevitable transformation of modern democratic practice through platform-dependent and transmedia political discourse. Political actors and media practitioners now deploy layers of social media network platforms to convey messages and mobilize supporters in order to aggregate and maximize the impact of their media campaign projects and audience reach.

Keywords: social media, digital humanities, political discourse, platformized discourse, multimodal discourse

Procedia PDF Downloads 57

Authors: Hyun-Koo Kim, Yong-Hun Kim, Soo-Young Suk, Ju H. Park, Ho-Youl Jung

Abstract:

This paper presents an effective speed hump detection process at the day-time. we focus only on round types of speed humps in the day-time dynamic road environment. The proposed speed hump detection scheme consists mainly of two process as stereo matching and speed hump detection process. Our proposed process focuses to speed hump detection process. Speed hump detection process consist of noise reduction step, data fusion step, and speed hemp detection step. The proposed system is tested on Intel Core CPU with 2.80 GHz and 4 GB RAM tested in the urban road environments. The frame rate of test videos is 30 frames per second and the size of each frame of grabbed image sequences is 1280 pixels by 670 pixels. Using object-marked sequences acquired with an on-vehicle camera, we recorded speed humps and non-speed humps samples. Result of the tests, our proposed method can be applied in real-time systems by computation time is 13 ms. For instance; our proposed method reaches 96.1 %.

Keywords: data fusion, round types speed hump, speed hump detection, surface filter

Procedia PDF Downloads 492

Authors: Wu You, Burra Venkata Durga Kumar

Abstract:

This study examines the promising potential of integrating Blockchain and Artificial Intelligence (AI) technologies to scalability in Distributed Systems within the field of supply chain finance. The finance industry is continually confronted with scalability challenges in its Distributed Systems, particularly within the supply chain finance sector, impacting efficiency and security. Blockchain, with its inherent attributes of high scalability and secure distributed ledger system, coupled with AI's strengths in optimizing data processing and decision-making, holds the key to innovating the industry's approach to these issues. This study elucidates the synergistic interplay between Blockchain and AI, detailing how their fusion can drive a significant transformation in the supply chain finance sector's Distributed Systems. It offers specific use-cases within this field to illustrate the practical implications and potential benefits of this technological convergence. The study also discusses future possibilities and current challenges in implementing this groundbreaking approach within the context of supply chain finance. It concludes that the intersection of Blockchain and AI could ignite a new epoch of enhanced efficiency, security, and transparency in the Distributed Systems of supply chain finance within the financial industry.

Keywords: blockchain, artificial intelligence (AI), scaled distributed systems, supply chain finance, efficiency and security

Procedia PDF Downloads 56

Authors: Yong-Hui Zheng

Abstract:

Viruses hijack host machineries to propagate and spread, which disrupts cellular homeostasis and activates various counteractive mechanisms. Infection of enveloped viruses is dependent on their fusion proteins, which bind to viral receptors to allow virus entry into cells. Fusion proteins are glycoproteins and expressed in the endoplasmic reticulum (ER) by hijacking the secretory pathway. Previously, we reported that Zaire ebolavirus (EBOV)-glycoprotein (GP) expression induces ER stress, and EBOV-GP is targeted by the calnexin cycle to macro-ER-phagy for degradation. We now report that expression of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2/SARS2)-spike (S) protein also causes ER stress, and its expression is strongly downregulated by mannosidase alpha class 1B member 1 (MAN1B1), a class I α-mannosidase from the ER. MAN1B1 co-localizes with SARS2-S in the ER, and its downregulation of SARS2-S is blocked by inhibitors targeting lysosomes and autophagy, but not proteasomes, indicating SARS2-S degradation by autolysosomes. Notably, the SARS2-S degradation does not require the core autophagy machinery including ATG3, ATG5, ATG7, and phosphatidylinositol 3-kinase catalytic subunit type 3 (PI3KC3)/vacuolar protein sorting 34 (VPS34), and instead, it requires Beclin 1 (BECN1), a core component in the PI3KC3 complex. In addition, MAN1B1 does not trigger SARS2-S polyubiquitination, and consistently, the SARS2-S degradation does not require the autophagy receptor sequestosome 1 (SQSTM1)/p62. MAN1B1 also downregulates EBOV-GP similarly, but this degradation does not require BECN1. Collectively, we conclude that MAN1B1 downregulates viral fusions by micro-ER-phagy, and importantly, we have identified BECN1-dependent and BECN1-independent mechanisms for micro-ER-phagy.

Keywords: Micro-ER-phagy, reticulophagy, fusion proteins, ER stress

Procedia PDF Downloads 50

Authors: Balaji Devanathan, Gokul G, Raveena Divya, Abhishek Yadav, Sudhir K.Gupta

Abstract:

Background: Age estimation is a common problem in administrative settings, medico legal cases, and among athletes competing in different sports. Age estimation is a problem in medico legal problems that arise in hospitals when there has been a criminal abortion, when consenting to surgery or a general physical examination, when there has been infanticide, impotence, sterility, etc. Medical imaging progress has benefited forensic anthropology in various ways, most notably in the area of determining bone age. An efficient method for researching the epiphyseal union and other differences in the body's bones and joints is multi-slice computed tomography. There isn't a significant database on Indians available. So to obtain an Indian based database author has performed this original study. Methodologies: The appearance and fusion of ossification centre of sternoclavicular joint is evaluated, and grades were assigned accordingly. Using MSCT scans, we examined the relationship between the age of the deceased and alterations in the sternoclavicular joint during the appearance and union in 500 instances, 327 men and 173 females, in the age range of 0 to 25 years. Results: According to our research in both the male and female groups, the ossification centre for the medial end of the clavicle first appeared between the ages of 18.5 and 17.1 respectively. The age range of the partial union was 20.4 and 20.2 years old. The earliest age of complete fusion was 23 years for males and 22 years for females. For fusion of their sternebrae into one, age range is 11–24 years for females and 17–24 years. The fusion of the third and fourth sternebrae was completed by 11 years. The fusions of the first and second and second and third sternebrae occur by the age of 17 years. Furthermore, correlation and reliability were carried out which yielded significant results. Conclusion: With numerous exceptions, the projected values are consistent with a large number of the previously developed age charts. These variations may be caused by the ethnic or regional heterogeneity in the ossification pattern among the population under study. The pattern of bone maturation did not significantly differ between the sexes, according to the study. The study's age range was 0 to 25 years, and for obvious reasons, the majority of the occurrences occurred in the last five years, or between 20 and 25 years of age. This resulted in a comparatively smaller study population for the 12–18 age group, where age estimate is crucial because of current legal requirements. It will require specialized PMCT research in this age range to produce population standard charts for age estimate. The medial end of the clavicle is one of several ossification foci that are being thoroughly investigated since they are challenging to assess with a traditional X-ray examination. Combining the two has been shown to be a valid result when it comes to raising the age beyond eighteen.

Keywords: age estimation, sternoclavicular joint, medial clavicle, computed tomography

Procedia PDF Downloads 20

Authors: Patsarawadee Paojinda, Waranya Imprasittichai, Sudaratana R. Krungkrai, Nirianne Marie Q. Palacpac, Toshihiro Horii, Jerapan Krungkrai

Abstract:

Fusion of the last two enzymes in the pyrimidine biosynthetic pathway in the inversed order by having COOH-terminal orotate phosphoribosyltransferase (OPRT) and NH2-terminal orotidine 5'-monophosphate decarboxylase (OMPDC), as OMPDC-OPRT, are described in many organisms. Here, we produced gene fusions of Plasmodium falciparum OMPDC-OPRT and expressed the bifunctional protein in Escherichia coli. The enzyme was purified to homogeneity using affinity and anion-exchange chromatography, exhibited enzymatic activities and functioned as a dimer. The activities, although unstable, can be stabilized by its substrate and product during purification and long-term storage. Furthermore, the enzyme expressed a perfect catalytic efficiency (kcat/Km). The kcat was selectively enhanced up to 3 orders of magnitude, while the Km was not much affected and remained at low µM levels when compared to the monofunctional enzymes. The fusion of the two enzymes, creating a “super-enzyme” with perfect catalytic power and more flexibility, reflects cryptic relationship of enzymatic reactivaties and metabolic functions on molecular evolution.

Keywords: bifunctional enzyme, orotate phosphoribosyltransferase, orotidine 5'-monophosphate decarboxylase, plasmodium falciparum

Procedia PDF Downloads 263