Search results for: deep q-network

Authors: Shubham Gogri, Lucia Florescu

Abstract:

Medical diagnosis and radiotherapy treatment planning using Computed Tomography (CT) rely on the quantitative accuracy and quality of the CT images. At the same time, requirements for CT imaging include reducing the radiation dose exposure to patients and minimizing scanning time. A solution to this is the sparse-view CT technique, based on a reduced number of projection views. This, however, introduces a new problem— the incomplete projection data results in lower quality of the reconstructed images. To tackle this issue, deep learning methods have been applied to enhance the quality of the sparse-view CT images. A first approach involved employing Mir-Net, a dedicated deep neural network designed for image enhancement. This showed promise, utilizing an intricate architecture comprising encoder and decoder networks, along with the incorporation of the Charbonnier Loss. However, this approach was computationally demanding. Subsequently, a specialized Generative Adversarial Network (GAN) architecture, rooted in the Pix2Pix framework, was implemented. This GAN framework involves a U-Net-based Generator and a Discriminator based on Convolutional Neural Networks. To bolster the GAN's performance, both Charbonnier and Wasserstein loss functions were introduced, collectively focusing on capturing minute details while ensuring training stability. The integration of the perceptual loss, calculated based on feature vectors extracted from the VGG16 network pretrained on the ImageNet dataset, further enhanced the network's ability to synthesize relevant images. A series of comprehensive experiments with clinical CT data were conducted, exploring various GAN loss functions, including Wasserstein, Charbonnier, and perceptual loss. The outcomes demonstrated significant image quality improvements, confirmed through pertinent metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) between the corrected images and the ground truth. Furthermore, learning curves and qualitative comparisons added evidence of the enhanced image quality and the network's increased stability, while preserving pixel value intensity. The experiments underscored the potential of deep learning frameworks in enhancing the visual interpretation of CT scans, achieving outcomes with SSIM values close to one and PSNR values reaching up to 76.

Keywords: generative adversarial networks, sparse view computed tomography, CT image correction, Mir-Net

Procedia PDF Downloads 162

Authors: E. E. Tereshatov, M. Yu. Boltoeva, V. Mazan, M. F. Volia, C. M. Folden III

Abstract:

Room temperature ionic liquids (RTILs) are a class of liquid organic salts with melting points below 20 °C that are considered to be environmentally friendly ‘designers’ solvents. Pure hydrophobic ILs are known to extract metallic species from aqueous solutions. The closest analogues of ionic liquids are deep eutectic solvents (DESs), which are a eutectic mixture of at least two compounds with a melting point lower than that of each individual component. DESs are acknowledged to be attractive for organic synthesis and metal processing. Thus, these non-volatile and less toxic compounds are of interest for critical metal extraction. The US Department of Energy and the European Commission consider indium as a key metal. Its chemical homologue, thallium, is also an important material for some applications and environmental safety. The aim of this work is to systematically investigate In and Tl extraction from aqueous solutions into pure fluorinated ILs and hydrophobic DESs. The dependence of the Tl extraction efficiency on the structure and composition of the ionic liquid ions, metal oxidation state, and initial metal and aqueous acid concentrations have been studied. The extraction efficiency of the TlXz3–z anionic species (where X = Cl– and/or Br–) is greater for ionic liquids with more hydrophobic cations. Unexpectedly high distribution ratios (> 103) of Tl(III) were determined even by applying a pure ionic liquid as receiving phase. An improved mathematical model based on ion exchange and ion pair formation mechanisms has been developed to describe the co-extraction of two different anionic species, and the relative contributions of each mechanism have been determined. The first evidence of indium extraction into new quaternary ammonium- and menthol-based hydrophobic DESs from hydrochloric and oxalic acid solutions with distribution ratios up to 103 will be provided. Data obtained allow us to interpret the mechanism of thallium and indium extraction into ILs and DESs media. The understanding of Tl and In chemical behavior in these new media is imperative for the further improvement of separation and purification of these elements.

Keywords: deep eutectic solvents, indium, ionic liquids, thallium

Procedia PDF Downloads 241

Authors: Hritwik Ghosh, Irfan Sadiq Rahat, Sachi Nandan Mohanty, J. V. R. Ravindra

Abstract:

In the rapidly evolving landscape of medical diagnostics, the early detection and accurate classification of skin cancer remain paramount for effective treatment outcomes. This research delves into the transformative potential of Artificial Intelligence (AI), specifically Deep Learning (DL), as a tool for discerning and categorizing various skin conditions. Utilizing a diverse dataset of 3,000 images representing nine distinct skin conditions, we confront the inherent challenge of class imbalance. This imbalance, where conditions like melanomas are over-represented, is addressed by incorporating class weights during the model training phase, ensuring an equitable representation of all conditions in the learning process. Our pioneering approach introduces a hybrid model, amalgamating the strengths of two renowned Convolutional Neural Networks (CNNs), VGG16 and ResNet50. These networks, pre-trained on the ImageNet dataset, are adept at extracting intricate features from images. By synergizing these models, our research aims to capture a holistic set of features, thereby bolstering classification performance. Preliminary findings underscore the hybrid model's superiority over individual models, showcasing its prowess in feature extraction and classification. Moreover, the research emphasizes the significance of rigorous data pre-processing, including image resizing, color normalization, and segmentation, in ensuring data quality and model reliability. In essence, this study illuminates the promising role of AI and DL in revolutionizing skin cancer diagnostics, offering insights into its potential applications in broader medical domains.

Keywords: artificial intelligence, machine learning, deep learning, skin cancer, dermatology, convolutional neural networks, image classification, computer vision, healthcare technology, cancer detection, medical imaging

Procedia PDF Downloads 87

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 157

Authors: Farina Riaz, Shahab Abdulla, Srinjoy Ganguly, Hajime Suzuki, Ravinesh C. Deo, Susan Hopkins

Abstract:

Recently, quantum machine learning has received significant attention. For various types of data, including text and images, numerous quantum machine learning (QML) models have been created and are being tested. Images are exceedingly complex data components that demand more processing power. Despite being mature, classical machine learning still has difficulties with big data applications. Furthermore, quantum technology has revolutionized how machine learning is thought of, by employing quantum features to address optimization issues. Since quantum hardware is currently extremely noisy, it is not practicable to run machine learning algorithms on it without risking the production of inaccurate results. To discover the advantages of quantum versus classical approaches, this research has concentrated on colored image data. Deep learning classification models are currently being created on Quantum platforms, but they are still in a very early stage. Black and white benchmark image datasets like MNIST and Fashion MINIST have been used in recent research. MNIST and CIFAR-10 were compared for binary classification, but the comparison showed that MNIST performed more accurately than colored CIFAR-10. This research will evaluate the performance of the QML algorithm on the colored benchmark dataset CIFAR-10 to advance QML's real-time applicability. However, deep learning classification models have not been developed to compare colored images like Quantum Convolutional Neural Network (QCNN) to determine how much it is better to classical. Only a few models, such as quantum variational circuits, take colored images. The methodology adopted in this research is a hybrid approach by using penny lane as a simulator. To process the 10 classes of CIFAR-10, the image data has been translated into grey scale and the 28 × 28-pixel image containing 10,000 test and 50,000 training images were used. The objective of this work is to determine how much the quantum approach can outperform a classical approach for a comprehensive dataset of color images. After pre-processing 50,000 images from a classical computer, the QCNN model adopted a hybrid method and encoded the images into a quantum simulator for feature extraction using quantum gate rotations. The measurements were carried out on the classical computer after the rotations were applied. According to the results, we note that the QCNN approach is ~12% more effective than the traditional classical CNN approaches and it is possible that applying data augmentation may increase the accuracy. This study has demonstrated that quantum machine and deep learning models can be relatively superior to the classical machine learning approaches in terms of their processing speed and accuracy when used to perform classification on colored classes.

Keywords: CIFAR-10, quantum convolutional neural networks, quantum deep learning, quantum machine learning

Procedia PDF Downloads 129

Authors: A. Beatti, L. Chipchase, A. Rayner, T. Souvlis

Abstract:

The aims of this study were to investigate the depth of interferential current (IFC) penetration through soft tissue and to investigate the area over which IFC spreads during clinical application. Premodulated IFC and ‘true’ IFC at beat frequencies of 4, 40 and 90Hz were applied via four electrodes to the distal medial thigh of 15 healthy subjects. The current was measured via three Teflon coated fine needle electrodes that were inserted into the superficial layer of skin, then into the subcutaneous tissue (≈1 cm deep) and then into muscle tissue (≈2 cm deep). The needle electrodes were placed in the middle of the four IFC electrodes, between two channels and outside the four electrodes. Readings were taken at each tissue depth from each electrode during each treatment frequency then digitized and stored for analysis. All voltages were greater at all depths and locations than baseline (p < 0.01) and voltages decreased with depth (P=0.039). Lower voltages of all currents were recorded in the middle of the four electrodes with the highest voltage being recorded outside the four electrodes in all depths (P=0.000).For each frequency of ‘true’ IFC, the voltage was higher in the superficial layer outside the electrodes (P ≤ 0.01).Premodulated had higher voltages along the line of one circuit (P ≤ 0.01). Clinically, IFC appears to pass through skin layers to depth and is more efficient than premodulated IFC when targeting muscle tissue.

Keywords: electrotherapy, interferential current, interferential therapy, medium frequency current

Procedia PDF Downloads 347

Authors: Merina Islam

Abstract:

There are many environmental problems. Various counter measures have been taken for environmental problems. Philosophy is an important contributor to environmental studies as it takes deep interest in meaning analysis of the concept environment and other related concepts. The Buddhist frame, which is virtue ethical, remains a better alternative to the traditional environmental outlook. Granting the unique role of man in immoral deliberations, the Buddhist approach, however, maintains a holistic concept of ecological harmony. Buddhist environmental ethics is more concerned about the complete moral community, the total ecosystem, than any particular species within the community. The moral reorientation proposed here has resemblance to the concept of 'deep ecology. Given the present day prominence of virtue ethics, we need to explore further into the Buddhist virtue theory, so that a better framework to treat the natural world would be ensured. Environment has turned out to be one of the most widely discussed issues in the recent times. Buddhist concepts such as Pratityasamutpadavada, Samvrit Satya, Paramartha Satya, Shunyata, Sanghatvada, Bodhisattva, Santanvada and others deal with interdependence in terms of both internal as well external ecology. The internal ecology aims at mental well-being whereas external ecology deals with physical well-being. The fundamental Buddhist concepts for dealing with environmental Problems are where the environment has the same value as humans as from the two Buddhist doctrines of the Non-duality of Life and its Environment and the Origination in Dependence; and the inevitability of overcoming environmental problems through the practice of the way of the Bodhisattva, because environmental problems are evil for people and nature. Buddhism establishes that there is a relationship among all the constituents of the world. There is nothing in the world which is independent from any other thing. Everything is dependent on others. The realization that everything in the universe is mutually interdependent also shows that the man cannot keep itself unaffected from ecology. This paper would like to focus how the Buddhist’s identification of nature and the Dhamma can contribute toward transforming our understanding, attitudes, and actions regarding the care of the earth. Environmental Ethics in Buddhism presents a logical and thorough examination of the metaphysical and ethical dimensions of early Buddhist literature. From the Buddhist viewpoint, humans are not in a category that is distinct and separate from other sentient beings, nor are they intrinsically superior. All sentient beings are considered to have the Buddha-nature, that is, the potential to become fully enlightened. Buddhists do not believe in treating of non-human sentient beings as objects for human consumption. The significance of Buddhist theory of interdependence can be understood from the fact that it shows that one’s happiness or suffering originates from ones realization or non-realization respectively of the dependent nature of everything. It is obvious, even without emphasis, which in the context of deep ecological crisis of today there is a need to infuse the consciousness of interdependence.

Keywords: Buddhism, deep ecology, environmental problems, Pratityasamutpadavada

Procedia PDF Downloads 315

Authors: Zakia Alioua, Amira Soumia, Zerouali-Khodja Fatiha

Abstract:

In spite of a wide variety of contaminants such as heavy metals and organic compounds in addition to the importance of extended pollution, the deep-sea and its species are not in haven and being affected through contaminants exposure. This investigation is performed in order to provide data on the presence of pathological changes in the liver and gonads of the greater forkbeard. A total of 998 specimens of the teleost fish Phycis blennoides Brünnich, 1768 ranged from 5,7 to 62,7 cm in total length, were obtained from the commercial fisheries of Algerian ports. The sampling has been carried out monthly from December 2013 to June 2015 and from January to June 2016 caught by trawlers and longlines between 75 and 600 fathoms in the coast of Algeria. Individuals were sexed their gonads, and their livers were removed and processed for light microscopy and one case of atresia was identified. In whole, overall 0,002% of the specimens presented some degree of liver steatosis. For the gastric section, 442 selected stomachs contents were observed looking for parasitic infestation and enumerate 212 nematodes. A prospecting survey for metal contaminant was performed on the liver by atomic absorption spectrophotometry analysis.

Keywords: atresia, coast of Algeria, histopathology, nematode, Phycis blennoides, steatosis

Procedia PDF Downloads 231

Authors: Lenka Hurtalová, Eva Tillová, Mária Chalupová, Juraj Belan, Milan Uhríčik

Abstract:

The secondary alloy A226 is used for many automotive casting produced by mould casting and high pressure die-casting. This alloy has excellent castability, good mechanical properties and cost-effectiveness. Production of primary aluminium alloys belong to heavy source fouling of life environs. The European Union calls for the emission reduction and reduction in energy consumption, therefore, increase production of recycled (secondary) aluminium cast alloys. The contribution is deal with influence of recycling on the quality of the casting made from A226 in automotive industry. The properties of the casting made from secondary aluminium alloys were compared with the required properties of primary aluminium alloys. The effect of recycling on microstructure was observed using combination different analytical techniques (light microscopy upon black-white etching, scanning electron microscopy-SEM upon deep etching and energy dispersive X-ray analysis-EDX). These techniques were used for the identification of the various structure parameters, which was used to compare secondary alloy microstructure with primary alloy microstructure.

Keywords: A226 secondary aluminium alloy, deep etching, mechanical properties, recycling foundry aluminium alloy

Procedia PDF Downloads 541

Authors: Nathenal Thomas

Abstract:

In the food chain and in many other agricultural products, corn, also known as maize, which goes by the scientific name Zea mays subsp, is a widely produced agricultural product. Corn has the highest adaptability. It comes in many different types, is employed in many different industrial processes, and is more adaptable to different agro-climatic situations. In Ethiopia, maize is among the most widely grown crop. Small-scale corn farming may be a household's only source of food in developing nations like Ethiopia. The aforementioned data demonstrates that the country's requirement for this crop is excessively high, and conversely, the crop's productivity is very low for a variety of reasons. The most damaging disease that greatly contributes to this imbalance between the crop's supply and demand is the corn disease. The failure to diagnose diseases in maize plant until they are too late is one of the most important factors influencing crop output in Ethiopia. This study will aid in the early detection of such diseases and support farmers during the cultivation process, directly affecting the amount of maize produced. The diseases in maize plants, such as northern leaf blight and cercospora leaf spot, have distinct symptoms that are visible. This study aims to detect the most frequent and degrading maize diseases using the most efficiently used subset of machine learning technology, deep learning so, called Image Processing. Deep learning uses networks that can be trained from unlabeled data without supervision (unsupervised). It is a feature that simulates the exercises the human brain goes through when digesting data. Its applications include speech recognition, language translation, object classification, and decision-making. Convolutional Neural Network (CNN) for Image Processing, also known as convent, is a deep learning class that is widely used for image classification, image detection, face recognition, and other problems. it will also use this algorithm as the state-of-the-art for my research to detect maize diseases by photographing maize leaves using a mobile phone.

Keywords: CNN, zea mays subsp, leaf blight, cercospora leaf spot

Procedia PDF Downloads 74

Authors: Nazmul Shahadat, Anthony S. Maida

Abstract:

In recent years, hypercomplex inspired neural networks improved deep CNN architectures due to their ability to share weights across input channels and thus improve cohesiveness of representations within the layers. The work described herein studies the effect of replacing existing layers in an Axial Attention ResNet with their quaternion variants that use cross-channel weight sharing to assess the effect on image classification. We expect the quaternion enhancements to produce improved feature maps with more interlinked representations. We experiment with the stem of the network, the bottleneck layer, and the fully connected backend by replacing them with quaternion versions. These modifications lead to novel architectures which yield improved accuracy performance on the ImageNet300k classification dataset. Our baseline networks for comparison were the original real-valued ResNet, the original quaternion-valued ResNet, and the Axial Attention ResNet. Since improvement was observed regardless of which part of the network was modified, there is a promise that this technique may be generally useful in improving classification accuracy for a large class of networks.

Keywords: axial attention, representational networks, weight sharing, cross-channel correlations, quaternion-enhanced axial attention, deep networks

Procedia PDF Downloads 83

Authors: Muhammad Zubair Khan, Yugyung Lee

Abstract:

Deep learning has recently achieved enormous response in semantic image segmentation. The previously developed U-Net inspired architectures operate with continuous stride and pooling operations, leading to spatial data loss. Also, the methods lack establishing long-term pixels connection to preserve context knowledge and reduce spatial loss in prediction. This article developed encoder-decoder architecture with bi-directional LSTM embedded in long skip-connections and densely connected convolution blocks. The network non-linearly combines the feature maps across encoder-decoder paths for finding dependency and correlation between image pixels. Additionally, the densely connected convolutional blocks are kept in the final encoding layer to reuse features and prevent redundant data sharing. The method applied batch-normalization for reducing internal covariate shift in data distributions. The empirical evidence shows a promising response to our method compared with other semantic segmentation techniques.

Keywords: deep learning, semantic segmentation, image analysis, pixels connection, convolution neural network

Procedia PDF Downloads 103

Authors: Dalia Ambrazaitienė, Monika Vilkienė, Danute Karcauskienė, Gintaras Siaudinis

Abstract:

The soil is a complex ecosystem that is part of our biosphere. The ability of soil to provide ecosystem services is dependent on microbial diversity. T Tillage is one of the major factors that affect soil properties. The no-till systems or shallow ploughless tillage are opposite of traditional deep ploughing, no-tillage systems, for instance, increase soil organic matter by reducing mineralization rates and stimulating litter concentrations of the top soil layer, whereas deep ploughing increases the biological activity of arable soil layer and reduces the incidence of weeds. The role of soil organisms is central to soil processes. Although the number of microbial species in soil is still being debated, the metagenomic approach to estimate microbial diversity predicted about 2000 – 18 000 bacterial genomes in 1 g of soil. Despite the key role of bacteria in soil processes, there is still lack of information about the bacterial diversity of soils as affected by tillage practices. This study focused on metagenomic analysis of bacterial diversity in long-term experimental plots of Dystric Epihypogleyic Albeluvisols in western part of Lithuania. The experiment was set up in 2013 and had a split-plot design where the whole-plot treatments were laid out in a randomized design with three replicates. The whole-plot treatments consisted of two tillage methods - deep ploughing (22-25 cm) (DP), ploughless tillage (7-10 cm) (PT). Three subsamples (0-20 cm) were collected on October 22, 2015 for each of the three replicates. Subsamples from the DP and PT systems were pooled together wise to make two composition samples, one representing deep ploughing (DP) and the other ploughless tillage (PT). Genomic DNA from soil sample was extracted from approximately 200 mg field-moist soil by using the D6005 Fungal/Bacterial Miniprep set (Zymo Research®) following the manufacturer’s instructions. To determine bacterial diversity and community composition, we employed a culture – independent approach of high-throughput pyrosequencing of the 16S rRNA gene. Metagenomic sequencing was made with Illumina MiSeq platform in Base Clear Company. The microbial component of soil plays a crucial role in cycling of nutrients in biosphere. Our study was a preliminary attempt at observing bacterial diversity in soil under two common but contrasting tillage practices. The number of sequenced reads obtained for PT (161 917) was higher than DP (131 194). The 10 most abundant genus in soil sample were the same (Arthrobacter, Candidatus Saccharibacteria, Actinobacteria, Acidobacterium, Mycobacterium, Bacillus, Alphaproteobacteria, Longilinea, Gemmatimonas, Solirubrobacter), just the percent of community part was different. In DP the Arthrobacter and Acidobacterium consist respectively 8.4 % and 2.5%, meanwhile in PT just 5.8% and 2.1% of all community. The Nocardioides and Terrabacter were observed just in PT. This work was supported by the project VP1-3.1-ŠMM-01-V-03-001 NKPDOKT and National Science Program: The effect of long-term, different-intensity management of resources on the soils of different genesis and on other components of the agro-ecosystems [grant number SIT-9/2015] funded by the Research Council of Lithuania.

Keywords: deep ploughing, metagenomics, ploughless tillage, soil community analysis

Procedia PDF Downloads 246

Authors: Kritsada Moonpa, Kannipa Motanated, Weerapan Srichan

Abstract:

Hong Hoi Formation is a Middle Triassic deep marine succession presented in outcrops throughout the Lampang Basin of northern Thailand. The primary goal of this research is to diagnose the paleoenvironment, petrographic compositions, and sedimentary sources of the Hong Hoi Formation in Ban Huat, Ngao District. The Triassic Hong Hoi Formation is chosen because the outcrops are continuous and fossils are greatly exposed and abundant. Depositional environment is reconstructed through sedimentological studies along with facies analysis. The Hong Hoi Formation is petrographically divided into two major facies, they are: sandstones with mudstone interbeds, and mudstones or shale with sandstone interbeds. Sandstone beds are lithic arenite and lithic greywacke, volcanic lithic fragments are dominated. Sedimentary structures, paleocurrent data and lithofacies arrangement indicate that the formation deposited in a part of deep marine abyssal plain environment. The sedimentological and petrographic features suggest that during the deposition the Hong Hoi Formation received sediment supply from nearby volcanic arc. This suggested that the intensive volcanic activity within the Sukhothai Arc during the Middle Triassic is the main sediment source.

Keywords: Sukhothai zone, petrography, Hong Hoi formation, Lampang, Triassic

Procedia PDF Downloads 213

Authors: Mahmut Yildirim

Abstract:

This paper proposed the bidirectional long short-term memory (Bi-LSTM) network-aided deep learning (DL)-based signal detection for Orthogonal frequency division multiplexing with all index modulation (OFDM-AIM), namely Bi-DeepAIM. OFDM-AIM is developed to increase the spectral efficiency of OFDM with index modulation (OFDM-IM), a promising multi-carrier technique for communication systems beyond 5G. In this paper, due to its strong classification ability, Bi-LSTM is considered an alternative to the maximum likelihood (ML) algorithm, which is used for signal detection in the classical OFDM-AIM scheme. The performance of the Bi-DeepAIM is compared with LSTM network-aided DL-based OFDM-AIM (DeepAIM) and classic OFDM-AIM that uses (ML)-based signal detection via BER performance and computational time criteria. Simulation results show that Bi-DeepAIM obtains better bit error rate (BER) performance than DeepAIM and lower computation time in signal detection than ML-AIM.

Keywords: bidirectional long short-term memory, deep learning, maximum likelihood, OFDM with all index modulation, signal detection

Procedia PDF Downloads 72

Authors: Abe Degale D., Cheng Jian

Abstract:

When creating automated video surveillance systems, violent action recognition is crucial. In recent years, hand-crafted feature detectors have been the primary method for achieving violence detection, such as the recognition of fighting activity. Researchers have also looked into learning-based representational models. On benchmark datasets created especially for the detection of violent sequences in sports and movies, these methods produced good accuracy results. The Hockey dataset's videos with surveillance camera motion present challenges for these algorithms for learning discriminating features. Image recognition and human activity detection challenges have shown success with deep representation-based methods. For the purpose of detecting violent images and identifying aggressive human behaviours, this research suggested a deep representation-based model using the transfer learning idea. The results show that the suggested approach outperforms state-of-the-art accuracy levels by learning the most discriminating features, attaining 99.34% and 99.98% accuracy levels on the Hockey and Movies datasets, respectively.

Keywords: violence detection, faster RCNN, transfer learning and, surveillance video

Procedia PDF Downloads 108

Authors: Lianzhong Zhang, Chao Huang

Abstract:

Sea-land segmentation is a basic step in many tasks such as sea surface monitoring and ship detection. The existing sea-land segmentation algorithms have poor segmentation accuracy, and the parameter adjustments are cumbersome and difficult to meet actual needs. Also, the current sea-land segmentation adopts traditional deep learning models that use Convolutional Neural Networks (CNN). At present, the transformer architecture has achieved great success in the field of natural images, but its application in the field of radar images is less studied. Therefore, this paper proposes a sea-land segmentation method based on the transformer architecture to strengthen edge supervision. It uses a self-attention mechanism with a gating strategy to better learn relative position bias. Meanwhile, an additional edge supervision branch is introduced. The decoder stage allows the feature information of the two branches to interact, thereby improving the edge precision of the sea-land segmentation. Based on the Gaofen-3 satellite image dataset, the experimental results show that the method proposed in this paper can effectively improve the accuracy of sea-land segmentation, especially the accuracy of sea-land edges. The mean IoU (Intersection over Union), edge precision, overall precision, and F1 scores respectively reach 96.36%, 84.54%, 99.74%, and 98.05%, which are superior to those of the mainstream segmentation models and have high practical application values.

Keywords: SAR, sea-land segmentation, deep learning, transformer

Procedia PDF Downloads 181

Authors: Jamshaid Ul Rahman, Faiza Makhdoom, Dianchen Lu

Abstract:

Activation functions play a decisive role in determining the capacity of Deep Neural Networks (DNNs) as they enable neural networks to capture inherent nonlinearities present in data fed to them. The prior research on activation functions primarily focused on the utility of monotonic or non-oscillatory functions, until Growing Cosine Unit (GCU) broke the taboo for a number of applications. In this paper, a Convolutional Neural Network (CNN) model named as ASU-CNN is proposed which utilizes recently designed activation function ASU across its layers. The effect of this non-monotonic and oscillatory function is inspected through feature map visualizations from different convolutional layers. The optimization of proposed network is offered by Adam with a fine-tuned adjustment of learning rate. The network achieved promising results on both training and testing data for the classification of CIFAR-10. The experimental results affirm the computational feasibility and efficacy of the proposed model for performing tasks related to the field of computer vision.

Keywords: amplifying sine unit, activation function, convolutional neural networks, oscillatory activation, image classification, CIFAR-10

Procedia PDF Downloads 111

Authors: Deawan Rakin Ahamed Remal, Sinthia Chowdhury, Sharun Akter Khushbu, Sheak Rashed Haider Noori

Abstract:

Native Speech to text synthesis has its own leverage for the purpose of mankind. The extensive nature of art to speaking different accents is common but the purpose of communication between two different accent types of people is quite difficult. This problem will be motivated by the extraction of the wrong perception of language meaning. Thus, many existing automatic speech recognition has been placed to detect text. Overall study of this paper mentions a review of NSTTR (Native Speech Text to Text Recognition) synthesis compared with Text to Text recognition. Review has exposed many text to text recognition systems that are at a very early stage to comply with the system by native speech recognition. Many discussions started about the progression of chatbots, linguistic theory another is rule based approach. In the Recent years Deep learning is an overwhelming chapter for text to text learning to detect language nature. To the best of our knowledge, In the sub continent a huge number of people speak in Bangla language but they have different accents in different regions therefore study has been elaborate contradictory discussion achievement of existing works and findings of future needs in Bangla language acoustic accent.

Keywords: TTR, NSTTR, text to text recognition, deep learning, natural language processing

Procedia PDF Downloads 132

Authors: Jacqueline Rose T. Alipo-on, Francesca Isabelle F. Escobar, Myles Joshua T. Tan, Hezerul Abdul Karim, Nouar Al Dahoul

Abstract:

Electrocardiogram (ECG) signal analysis and processing are crucial in the diagnosis of cardiovascular diseases, which are considered one of the leading causes of mortality worldwide. However, the traditional rule-based analysis of large volumes of ECG data is time-consuming, labor-intensive, and prone to human errors. With the advancement of the programming paradigm, algorithms such as machine learning have been increasingly used to perform an analysis of ECG signals. In this paper, various deep learning algorithms were adapted to classify five classes of heartbeat types. The dataset used in this work is the synthetic MIT-BIH Arrhythmia dataset produced from generative adversarial networks (GANs). Various deep learning models such as ResNet-50 convolutional neural network (CNN), 1-D CNN, and long short-term memory (LSTM) were evaluated and compared. ResNet-50 was found to outperform other models in terms of recall and F1 score using a five-fold average score of 98.88% and 98.87%, respectively. 1-D CNN, on the other hand, was found to have the highest average precision of 98.93%.

Keywords: heartbeat classification, convolutional neural network, electrocardiogram signals, generative adversarial networks, long short-term memory, ResNet-50

Procedia PDF Downloads 128

Authors: Akshay Paranjape, Nils Plettenberg, Robert Schmitt

Abstract:

Traditional optimization methods like evolutionary algorithms are widely used in production processes to find an optimal or near-optimal solution of control parameters based on the simulated environment space of a process. These algorithms are computationally intensive and therefore do not provide the opportunity for real-time optimization. This paper utilizes the Deep Reinforcement Learning (DRL) framework to find an optimal or near-optimal solution for control parameters. A model based on maximum a posteriori policy optimization (Hybrid-MPO) that can handle both numerical and categorical parameters is used as a benchmark for comparison. A comparative study shows that DRL can find optimal solutions of similar quality as compared to evolutionary algorithms while requiring significantly less time making them preferable for real-time optimization. The results are confirmed in a large-scale validation study on datasets from production and other fields. A trained XGBoost model is used as a surrogate for process simulation. Finally, multiple ways to improve the model are discussed.

Keywords: reinforcement learning, evolutionary algorithms, production process optimization, real-time optimization, hybrid-MPO

Procedia PDF Downloads 112

Authors: Niousha Bagheri Khulenjani, Mohammad Saniee Abadeh

Abstract:

Learning from very big datasets is a significant problem for most present data mining and machine learning algorithms. MicroRNA (miRNA) is one of the important big genomic and non-coding datasets presenting the genome sequences. In this paper, a hybrid method for the classification of the miRNA data is proposed. Due to the variety of cancers and high number of genes, analyzing the miRNA dataset has been a challenging problem for researchers. The number of features corresponding to the number of samples is high and the data suffer from being imbalanced. The feature selection method has been used to select features having more ability to distinguish classes and eliminating obscures features. Afterward, a Convolutional Neural Network (CNN) classifier for classification of cancer types is utilized, which employs a Genetic Algorithm to highlight optimized hyper-parameters of CNN. In order to make the process of classification by CNN faster, Graphics Processing Unit (GPU) is recommended for calculating the mathematic equation in a parallel way. The proposed method is tested on a real-world dataset with 8,129 patients, 29 different types of tumors, and 1,046 miRNA biomarkers, taken from The Cancer Genome Atlas (TCGA) database.

Keywords: cancer classification, feature selection, deep learning, genetic algorithm

Procedia PDF Downloads 111

Authors: Dimassi Wissem

Abstract:

In this work, we investigate the effect of combination between the porous silicon (PS) layer passivized with Lithium Bromide (LiBr) and grooving of grain boundaries (GB) in multi crystalline silicon. The grain boundaries were grooved in order to reduce the area of these highly recombining regions. Using optimized conditions, grooved GB's enable deep phosphorus diffusion and deep metallic contacts. We have evaluated the effects of LiBr on the surface properties of porous silicon on the performance of silicon solar cells. The results show a significant improvement of the internal quantum efficiency, which is strongly related to the photo-generated current. We have also shown a reduction of the surface recombination velocity and an improvement of the diffusion length after the LiBr process. As a result, the I–V characteristics under the dark and AM1.5 illumination were improved. It was also observed a reduction of the GB recombination velocity, which was deduced from light-beam-induced-current (LBIC) measurements. Such grooving in multi crystalline silicon enables passivization of GB-related defects. These results are discussed and compared to solar cells based on untreated multi crystalline silicon wafers.

Keywords: Multicrystalline silicon, LiBr, porous silicon, passivation

Procedia PDF Downloads 396

Authors: Hiroyuki Aoki

Abstract:

Molecular imaging has attracted much attention to visualize a tumor site in a living body on the basis of biological functions. A fluorescence in vivo imaging technique has been widely employed as a useful modality for small animals in pre-clinical researches. However, it is difficult to observe a site deep inside a body because of a short penetration depth of light. A photo-acoustic effect is a generation of a sound wave following light absorption. Because the sound wave is less susceptible to the absorption of tissues, an in vivo imaging method based on the photoacoustic effect can observe deep inside a living body. The current study developed an in vivo imaging agent for a photoacoustic imaging method. Nano-particles of poly(lactic acid) including indocyanine dye were developed as bio-compatible imaging agent with strong light absorption. A tumor site inside a mouse body was successfully observed in a photo-acoustic image. A photo-acoustic imaging with polymer nano-particle agent would be a powerful method to visualize a tumor.

Keywords: nano-particle, photo-acoustic effect, polymer, dye, in vivo imaging

Procedia PDF Downloads 155

Authors: Mo-Se Lee, Cheol-Hwi Ahn, Kee-Young Kwahk, Hyunchul Ahn

Abstract:

Over the past decade, deep learning has been in spotlight among various machine learning algorithms. In particular, CNN (Convolutional Neural Network), which is known as effective solution for recognizing and classifying images, has been popularly applied to classification and prediction problems in various fields. In this study, we try to apply CNN to stock market prediction, one of the most challenging tasks in the machine learning research. In specific, we propose to apply CNN as the binary classifier that predicts stock market direction (up or down) by using a graph as its input. That is, our proposal is to build a machine learning algorithm that mimics a person who looks at the graph and predicts whether the trend will go up or down. Our proposed model consists of four steps. In the first step, it divides the dataset into 5 days, 10 days, 15 days, and 20 days. And then, it creates graphs for each interval in step 2. In the next step, CNN classifiers are trained using the graphs generated in the previous step. In step 4, it optimizes the hyper parameters of the trained model by using the validation dataset. To validate our model, we will apply it to the prediction of KOSPI200 for 1,986 days in eight years (from 2009 to 2016). The experimental dataset will include 14 technical indicators such as CCI, Momentum, ROC and daily closing price of KOSPI200 of Korean stock market.

Keywords: convolutional neural network, deep learning, Korean stock market, stock market prediction

Procedia PDF Downloads 425

Authors: A. Degale Desta, Tamirat Kebamo

Abstract:

Background: Recent advances in computer vision and pattern recognition have significantly improved activity recognition through video analysis, particularly with the application of Deep Convolutional Neural Networks (CNNs). One-stage detectors now enable efficient video-based recognition by simultaneously predicting object categories and locations. Such advancements are highly relevant in educational settings where CCTV surveillance could automatically monitor academic activities, enhancing security and classroom management. However, current datasets and recognition systems lack the specific focus on campus environments necessary for practical application in these settings.Objective: This study aims to address this gap by developing a dataset and testing an automated activity recognition system specifically tailored for educational campuses. The EthioCAD dataset was created to capture various classroom activities and teacher-student interactions, facilitating reliable recognition of academic activities using deep learning models. Method: EthioCAD, a novel video-based dataset, was created with a design science research approach to encompass teacher-student interactions across three domains and 18 distinct classroom activities. Using the Roboflow AI framework, the data was processed, with 4.224 KB of frames and 33.485 MB of images managed for frame extraction, labeling, and organization. The Ultralytics YOLOv8 model was then implemented within Google Colab to evaluate the dataset’s effectiveness, achieving high mean Average Precision (mAP) scores. Results: The YOLOv8 model demonstrated robust activity recognition within campus-like settings, achieving an mAP50 of 90.2% and an mAP50-95 of 78.6%. These results highlight the potential of EthioCAD, combined with YOLOv8, to provide reliable detection and classification of classroom activities, supporting automated surveillance needs on educational campuses. Discussion: The high performance of YOLOv8 on the EthioCAD dataset suggests that automated activity recognition for surveillance is feasible within educational environments. This system addresses current limitations in campus-specific data and tools, offering a tailored solution for academic monitoring that could enhance the effectiveness of CCTV systems in these settings. Conclusion: The EthioCAD dataset, alongside the YOLOv8 model, provides a promising framework for automated campus activity recognition. This approach lays the groundwork for future advancements in CCTV-based educational surveillance systems, enabling more refined and reliable monitoring of classroom activities.

Keywords: deep CNN, EthioCAD, deep learning, YOLOv8, activity recognition

Procedia PDF Downloads 12

Authors: Hani Mekdash, Lina Jaber, Yehia Temsah

Abstract:

Underground space is becoming a necessity nowadays, especially in highly congested urban areas. Retaining underground excavations using shoring systems is essential in order to protect adjoining structures from potential damage or collapse. Reinforced Concrete Piles (RCP) supported by multiple rows of tie-back anchors are commonly used type of shoring systems in deep excavations. However, executing anchors can sometimes be challenging because they might illegally trespass neighboring properties or get obstructed by infrastructure and other underground facilities. A technique is proposed in this paper, and it involves the addition of eccentric high-strength steel strands to the RCP section through ducts without providing the pile with lateral supports. The strands are then vertically stressed externally on the pile cap using a hydraulic jack, creating a compressive strengthening force in the concrete section. An experimental study about the behavior of the shoring wall by pre-stressed piles is presented during the execution of an open excavation in an urban area (Beirut city) followed by numerical analysis using finite element software. Based on the experimental results, this technique is proven to be cost-effective and provides flexible and sustainable construction of shoring works.

Keywords: deep excavation, prestressing, pre-stressed piles, shoring system

Procedia PDF Downloads 117

Authors: Kritiyaporn Kunsook

Abstract:

Machine learning algorithms (MLAs) such us artificial neural networks (ANNs), decision tree, support vector machines (SVMs), Naïve Bayes, and ensemble classifier by voting are powerful data driven methods that are relatively less widely used in the mapping of technique of system, and thus have not been comparatively evaluated together thoroughly in this field. The performances of a series of MLAs, ANNs, decision tree, SVMs, Naïve Bayes, and ensemble classifier by voting in technique of hydroponic systems prospectively modeling are compared based on the accuracy of each model. Classification of hydroponic systems only covers the test samples from vegetables grown with Nutrient film technique (NFT) and Deep flow technique (DFT). The feature, which are the characteristics of vegetables compose harvesting height width, temperature, require light and color. The results indicate that the classification performance of the ANNs is 98%, decision tree is 98%, SVMs is 97.33%, Naïve Bayes is 96.67%, and ensemble classifier by voting is 98.96% algorithm respectively.

Keywords: artificial neural networks, decision tree, support vector machines, naïve Bayes, ensemble classifier by voting

Procedia PDF Downloads 372

Authors: Fani Sakellariadou

Abstract:

The Blue Economy includes all marine-based and marine-related activities. They correspond to established, emerging as well as unborn ocean-based industries. Seabed mining is an emerging marine-based activity; its operations depend particularly on cutting-edge science and technology. The 21st century will face a crisis in resources as a consequence of the world’s population growth and the rising standard of living. The natural capital stored in the global ocean is decisive for it to provide a wide range of sustainable ecosystem services. Seabed mineral deposits were identified as having a high potential for critical elements and base metals. They have a crucial role in the fast evolution of green technologies. The major categories of marine mineral deposits are deep-sea deposits, including cobalt-rich ferromanganese crusts, polymetallic nodules, phosphorites, and deep-sea muds, as well as shallow-water deposits including marine placers. Seabed mining operations may take place within continental shelf areas of nation-states. In international waters, the International Seabed Authority (ISA) has entered into 15-year contracts for deep-seabed exploration with 21 contractors. These contracts are for polymetallic nodules (18 contracts), polymetallic sulfides (7 contracts), and cobalt-rich ferromanganese crusts (5 contracts). Exploration areas are located in the Clarion-Clipperton Zone, the Indian Ocean, the Mid Atlantic Ridge, the South Atlantic Ocean, and the Pacific Ocean. Potential environmental impacts of deep-sea mining include habitat alteration, sediment disturbance, plume discharge, toxic compounds release, light and noise generation, and air emissions. They could cause burial and smothering of benthic species, health problems for marine species, biodiversity loss, reduced photosynthetic mechanism, behavior change and masking acoustic communication for mammals and fish, heavy metals bioaccumulation up the food web, decrease of the content of dissolved oxygen, and climate change. An important concern related to deep-sea mining is our knowledge gap regarding deep-sea bio-communities. The ecological consequences that will be caused in the remote, unique, fragile, and little-understood deep-sea ecosystems and inhabitants are still largely unknown. The blue economy conceptualizes oceans as developing spaces supplying socio-economic benefits for current and future generations but also protecting, supporting, and restoring biodiversity and ecological productivity. In that sense, people should apply holistic management and make an assessment of marine mining impacts on ecosystem services, including the categories of provisioning, regulating, supporting, and cultural services. The variety in environmental parameters, the range in sea depth, the diversity in the characteristics of marine species, and the possible proximity to other existing maritime industries cause a span of marine mining impact the ability of ecosystems to support people and nature. In conclusion, the use of the untapped potential of the global ocean demands a liable and sustainable attitude. Moreover, there is a need to change our lifestyle and move beyond the philosophy of single-use. Living in a throw-away society based on a linear approach to resource consumption, humans are putting too much pressure on the natural environment. Applying modern, sustainable and eco-friendly approaches according to the principle of circular economy, a substantial amount of natural resource savings will be achieved. Acknowledgement: This work is part of the MAREE project, financially supported by the Division VI of IUPAC. This work has been partly supported by the University of Piraeus Research Center.

Keywords: blue economy, deep-sea mining, ecosystem services, environmental impacts

Procedia PDF Downloads 83

Authors: Abdolreza Heidari Noosh Abad, Amir Abedi, Davood Mirahmadi khaki

Abstract:

The high strength and flexibility of TRIP steels are the major reasons for them being widely used in the automobile industry. Deep drawing is regarded as a common metal sheet manufacturing process is used extensively in the modern industry, particularly automobile industry. To investigate the potential of deep drawing characteristic of materials, steel sheet anisotropy is studied and expressed as R-Value. The TRIP steels have a multi-phase microstructure consisting typically of ferrite, bainite and retained austenite. The retained austenite appears to be the most effective phase in the microstructure of the TRIP steels. In the present research, Taguchi method has been employed to study investigates the effect of austempering temperature parameters on the anisotropy property of the TRIP steel. To achieve this purpose, a steel with chemical composition of 0.196C -1.42Si-1.41Mn, has been used and annealed at 810oC, and then austempered at 340-460oC for 3, 6, and 9 minutes. The results shows that the austempering temperature has a direct relationship with R-value, respectively. With increasing austempering temperature, residual austenite grain size increases as well as increased solubility, which increases the amount of R-value. According to the results of the Taguchi method, austempering temperature’s p-value less than 0.05 is due to effective on R-value.

Keywords: Taguchi method, hot rolling, thermomechanical process, anisotropy, R-value

Procedia PDF Downloads 326