REVIEW | doi:10.20944/preprints202311.1366.v1
Subject: Engineering, Bioengineering Keywords: transformers; artificial intelligence; GAN; medical images
Online: 22 November 2023 (03:41:01 CET)
The integration of artificial intelligence (AI) into medical imaging has guided in an era of transformation in healthcare. This literature review explores the latest innovations and applications of AI in the field, highlighting its profound impact on medical diagnosis and patient care. The innovation segment explores into cutting-edge developments in AI, such as deep learning algorithms, convolutional neural networks, and generative adversarial networks, which have significantly improved the accuracy and efficiency of medical image analysis. These innovations have enabled rapid and accurate detection of abnormalities, from identifying tumors during radiological examinations to detecting early signs of eye disease in retinal images. The article also highlights various applications of AI in medical imaging, including radiology, pathology, cardiology, and more. AI-based diagnostic tools not only speed up the interpretation of complex images but also improve early detection of disease, ultimately delivering better outcomes for patients. Additionally, AI-based image processing facilitates personalized treatment plans, thereby optimizing healthcare delivery. This literature review highlights the paradigm shift that AI has brought to medical imaging, highlighting its role in revolutionizing diagnosis and patient care. By combining cutting-edge AI techniques and their practical applications, it is clear that AI will continue shaping the future of healthcare in profound and positive ways.
ARTICLE | doi:10.20944/preprints202310.2053.v1
Subject: Chemistry And Materials Science, Electronic, Optical And Magnetic Materials Keywords: µLED; GaN; KOH treatment; sidewall damage
Online: 31 October 2023 (09:54:40 CET)
InGaN-based red micro light-emitting diodes (µLEDs) of different sizes prepared in this work. The red GaN epilayers were grown on a 4-inch sapphire substrate through metal organic chemical vapor deposition (MOCVD). Etching, sidewall treatment, and p- and n- contact deposition was involved in the fabrication process. Initially, the etching process would cause undesirable damages to the GaN sidewalls, which leads to an increase in leakage current. Hence, we employed KOH wet treatment to rectify the defects on the sidewalls and conducted a comparative and systematic analysis on electrical as well as optical properties. We observed that the µLEDs with a size of 5µm exhibited substantial leakage current, which was effectively mitigated by the application of KOH wet treatment. In terms of optical performance, the arrays with KOH demonstrated improved Light Output Power (LOP). Additionally, while photoelectric performance exhibited a decline with increasing current density, the devices treated with KOH consistently outperformed their counterparts in terms of optoelectronic efficiency. It is noteworthy that the optimized devices displayed enhanced photoelectric characteristics without significantly altering their original peak wavelength and FWHM. Our findings point to the elimination of surface non-radiative recombination by KOH wet treatment, thereby enhancing the performance of small-sized red µLEDs which has significant potential in realizing full-color micro-displays in near-eye projection applications.
ARTICLE | doi:10.20944/preprints202301.0174.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: GaN; MOSHFET; Ga2O3; MOCVD; Gate Dielectric
Online: 10 January 2023 (04:46:24 CET)
We report the electrical properties of Al0.3Ga0.7N/GaN heterojunction field effect transistor (HFET) structures with a Ga2O3 passivation layer grown by metal-organic chemical vapor deposition (MOCVD). In this study, three different thicknesses of Ga2O3 dielectric layers were grown on Al0.3Ga0.7N/GaN structures leading to metal-oxide-semiconductor-HFET or MOSHFET structures. X-ray diffraction (XRD) showed the (¯201) orientation peaks of -Ga2O3 in the device structure. The van der Pauw and Hall measurements yield the electron density of ~ 4 x1018 cm-3 and mobility of ~770 cm2V-1s-1 in the 2-dimensional electron gas (2DEG) channel at room temperature. Capacitance-voltage (C-V) measurement for the on-state 2DEG density for the MOSHFET structure was found to be of the order of ~1.5 x1013 cm-2. The thickness of the Ga2O3 layer was inversely related to the threshold voltage and the on-state capacitance. The interface charge density between the oxide and Al0.3Ga0.7N barrier layer was found to be of the order of ~1012 cm2eV-1. The annealing at 900° C of the MOSHFET structures revealed that the Ga2O3 layer was thermally stable at high temperatures resulting in insignificant threshold voltage shifts for annealed samples with respect to as-deposited (unannealed) structures. Our results show that the MOCVD-gown Ga2O3 dielectric layers can be a strong candidate for stable high-power devices.
ARTICLE | doi:10.20944/preprints202311.1723.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: magnetic bearings; MOSFETs; gallium nitride; GaN
Online: 27 November 2023 (16:12:47 CET)
Active magnetic bearings have higher efficiency than other bearings because there is no physical contact. However, this benefit is mitigated by the addition of electrical power consumption. It is therefore important for magnetic bearings to have efficient power electronics. Gallium nitride-based transistors are a relatively new form of transistor which have shown to be more efficient than MOSFETs. There is an insufficient body of literature in the area of application of these transistors for magnetic levitation. This work presents a simple 1 degree-of-freedom voltage controlled levitation experiment in which levitation is achieved with both a gallium nitride transistor and a MOSFET.Voltage losses and current consumption are measured when using each transistor during levitation. In particular, transients during open-to-closed and closed-to-open states are measured for PWM pulsing. It is found that thegallium nitride transistor is superior in both switching time and efficiency in situ for magnetic levitation.
ARTICLE | doi:10.20944/preprints202310.1290.v1
Subject: Computer Science And Mathematics, Computer Vision And Graphics Keywords: traditional oriental landscape (TOL); GAN; SPADE; generator; discriminator
Online: 19 October 2023 (16:44:14 CEST)
We present TOLGAN that generates traditional oriental landscape (TOL) image from a map that specifies the locations and shapes of the elements composing TOL. Users can create a TOL map by using a user interface or a segmentation scheme from a photograph. We design the generator of TOLGAN as a series of decoding layers where the map is applied between the layers. The generated TOL image is further enhanced through an AdaIN architecture. The discriminator of TOLGAN processes a generated image and its groundtruth TOL artwork image. TOLGAN is trained through a dataset composed of paired TOL artwork images and their TOL maps. We present a tool through which users can produce a TOL map by specifying and organizing the elements of TOL artworks. TOLGAN successfully generates a series of TOL images from the TOL map. We evaluate our approach using a quantitative way by estimating FID and ArtFID scores and a qualitative way by executing two user studies. Through these studies, we prove the excellence of our approach by comparing our results with those from several important existing works.
ARTICLE | doi:10.20944/preprints202304.0935.v1
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: Multilingual BERT; Text-to-image; DCGAN; Webtoon; GAN
Online: 26 April 2023 (03:16:07 CEST)
Recent advances in deep learning technology have led to increased interest in text-to-image technology, which enables computers to create images from text by simulating the human process of forming mental images. The GAN-based text-to-image technology involves the extraction of features from input text, which are combined with noise and then used as input to a GAN that generates images that are similar to the original images through competition between the generator and discriminator. Although generating images from English text is a mature area of research, text-to-image technology based on multilingualism, such as Korean, is still in its early stages of development. Webtoon is a digital comic format that allows comics to be viewed online. The creation process for webtoons is divided into story planning, content/sketching, coloring, and background drawing. Since each stage of webtoon production requires human intervention, it is both time-consuming and expensive. As a result, deep learning technologies such as automatic coloring and automatic line drawing are being used to reduce human involvement. However, there is a shortage of technology that can assist authors with story creation in webtoon production. Therefore, this study proposes a multilingual text-to-image model capable of generating webtoon images when presented with multilingual input text. The proposed model employs Multilingual BERT to extract feature vectors for multiple languages, and trains a DCGAN in conjunction with the images. The experimental results demonstrate that the model can generate images that are similar to the original images when presented with multilingual input text after training.
ARTICLE | doi:10.20944/preprints202012.0678.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: Ge; germanium; doping; GaN; gallium; nitride; MOVPE; epitaxy
Online: 28 December 2020 (10:40:29 CET)
The effect of growth temperature and precursor flows on the doping level and surface morphology of Ge-doped GaN layers was researched. The results show that germanium is more readily incorporated at low temperature, high growth rate and high V/III ratio, thus revealing a similar behavior to what was previously observed for indium. V-pit formation can be blocked at high temperature but also at low V/III ratio, the latter of which however causing step bunching.
Subject: Computer Science And Mathematics, Data Structures, Algorithms And Complexity Keywords: GAN; ECG; anonymization; healthcare data; sensors; data transformation
Online: 3 September 2020 (05:26:01 CEST)
In personalized healthcare, an ecosystem for the manipulation of reliable and safe private data should be orchestrated. This paper describes a first approach for the generation of fake electrocardiograms (ECGs) based on Generative Adversarial Networks (GANs) with the objective of anonymizing users’ information for privacy issues. This is intended to create valuable data that can be used both, in educational and research areas, while avoiding the risk of a sensitive data leakage. As GANs are mainly exploited on images and video frames, we are proposing general raw data processing after transformation into an image, so it can be managed through a GAN, then decoded back to the original data domain. The feasibility of our transformation and processing hypothesis is primarily demonstrated. Next, from the proposed procedure, main drawbacks for each step in the procedure are addressed for the particular case of ECGs. Hence, a novel research pathway on health data anonymization using GANs is opened and further straightforward developments are expected.
ARTICLE | doi:10.3390/sci2010013
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: deep learning; computer vision; Cycle-GAN; image reconstruction
Online: 12 March 2020 (00:00:00 CET)
In this paper, our goal is to perform a virtual restoration of an ancient coin from its image. The present work is the first one to propose this problem, and it is motivated by two key promising applications. The first of these emerges from the recently recognised dependence of automatic image based coin type matching on the condition of the imaged coins; the algorithm introduced herein could be used as a pre-processing step, aimed at overcoming the aforementioned weakness. The second application concerns the utility both to professional and hobby numismatists of being able to visualise and study an ancient coin in a state closer to its original (minted) appearance. To address the conceptual problem at hand, we introduce a framework which comprises a deep learning based method using Generative Adversarial Networks, capable of learning the range of appearance variation of different semantic elements artistically depicted on coins, and a complementary algorithm used to collect, correctly label, and prepare for processing a large numbers of images (here 100,000) of ancient coins needed to facilitate the training of the aforementioned learning method. Empirical evaluation performed on a withheld subset of the data demonstrates extremely promising performance of the proposed methodology and shows that our algorithm correctly learns the spectra of appearance variation across different semantic elements, and despite the enormous variability present reconstructs the missing (damaged) detail while matching the surrounding semantic content and artistic style.
ARTICLE | doi:10.20944/preprints202307.1860.v1
Subject: Physical Sciences, Condensed Matter Physics Keywords: Interfaces; AlN; GaN; pseudomorphic heterostructures; strain-balance; DFT; HRTEM
Online: 28 July 2023 (10:40:22 CEST)
We present the results of a twofold experimental and computational study of (0001) GaN/AlN multilayers forming pseudomorphic superlattices. High-Resolution Transmission Electron Microscopy (HRTEM) shows that heterostructures with GaN Quantum Wells (QW) four c-lattice parameters thick are misfit-dislocation free. Accurate structural data are extracted from HRTEM images via a new methodology optimizing the residual elastic energy stored in the samples. Total energy calculations are performed with several models analogous to the experimental QWs with increasing thicknesses of GaN, whereas this of the AlN barrier is kept fixed at n=8 c-lattice parameters. With vanishing external stresses, minimum energy configurations of the studied systems correspond to different strain states. Linear elasticity accurately yields the corresponding lattice parameters, suppressing the need for on-purpose total energy calculations. Theoretically justified parabolic fits of the excess interfacial energy yield the values of interfacial stress and elastic stiffness as functions of the GaN QW thickness. Total species-projected densities of states and gap values extracted therefrom allow deciphering the effect of the evolving strain on the electronic structure of the superlattice. It is found that the gap energy decreases linearly with increasing the strain of the QW. These results are briefly discussed in the light shed by previous works from the literature.
REVIEW | doi:10.20944/preprints202304.0764.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: GaN; micro-LED; non-radiative recombination; EQE; size effect
Online: 23 April 2023 (04:20:54 CEST)
GaN-based micro-size light-emitting diodes (µLEDs) have a number of appealing and distinctive benefits for display, visible-light communication (VLC), and other novel applications. The smaller size of LEDs affords them the benefits of enhanced current expansion, less self-heating effects, and higher current density bearing capacity. Low external quantum efficiency (EQE) resulting from non-radiative recombination and quantum confined stark effect (QCSE) is the serious barrier for applications of µLEDs. In this work, the reasons for the poor EQE of µLEDs are reviewed, as well as the optimization techniques for improving the EQE of µLEDs.
Subject: Engineering, Electrical And Electronic Engineering Keywords: forgery detection; GAN; contrastive loss; deep learning; pairwise learning
Online: 5 May 2019 (11:13:55 CEST)
Recently, generative adversarial networks (GANs) can be used to generate the photo-realistic image from a low-dimension random noise. It is very dangerous that the synthesized or generated image is used on inappropriate contents in social media network. In order to successfully detect such fake image, an effective and efficient image forgery detector is desired. However, conventional image forgery detectors are failed to recognize the synthesized or generated images by using GAN-based generator since they are all generated but manipulation from the source. Therefore, we propose a deep learning-based approach to detect the fake image by combining the contrastive loss. First, several state-of-the-art GANs will be collected to generate the fake-real image pairs. Then, the contrastive will be used on the proposed common fake feature network (CFFN)to learn the discriminative feature between the fake image and real image (i.e., paired information). Finally, a smaller network will be concatenated to the CFFN to determine whether the feature of the input image is fake or real. Experimental results demonstrated that the proposed method significantly outperforms other state-of-the-art fake image detectors.
ARTICLE | doi:10.20944/preprints202305.0705.v1
Subject: Medicine And Pharmacology, Oncology And Oncogenics Keywords: MR synthesis, GAN, multi-modal fusion, tumor monitoring, contrast enhancement
Online: 10 May 2023 (08:10:11 CEST)
Purposes: To provide abdominal contrast-enhanced MR image synthesis, we developed an image gradient regularized multi-modal multi-discrimination sparse-attention fusion generative adversarial network (GRMM-GAN) to avoid repeated contrast injections to patients and facilitate adaptive monitoring. Methods: With IRB approval, 165 abdominal MR studies from 61 liver cancer patients were retrospectively solicited from our institutional database. Each study included T2, T1 pre-contrast (T1pre), and T1 contrast-enhanced (T1ce) images. The GRMM-GAN synthesis pipeline consists of a sparse attention fusion network, an image gradient regularizer (GR), and a generative adversarial network with multi-discrimination. The studies were randomly divided into 115 for training, 20 for validation, and 30 for testing. The two pre-contrast MR modalities, T2 and T1pre images, were adopted as inputs in the training phase. The T1ce image at the portal venous phase was used as an output. The synthesized T1ce images were compared with the ground truth T1ce images. The evaluation metrics include peak signal-to-noise ratio (PSNR), structural-similarity-index (SSIM), and mean-squared-error (MSE). A Turing test and experts’ contours evaluated the image synthesis quality. Results: The proposed GRMM-GAN model achieved a PSNR of 28.56, an SSIM of 0.869, and an MSE of 83.27. The proposed model showed statistically significant improvements in all metrics tested with p-values <0.05 over the state-of-the-art model comparisons. The average Turing test score was 52.33%, close to random guessing, supporting the model's effectiveness for clinical application. In the tumor-specific region analysis, the average tumor contrast-to-noise ratio (CNR) of the synthesized MR images was not statistically significant from the real MR images. The average DICE from real vs. synthetic images was 0.90, compared to the inter-operator DICE of 0.91. Conclusion: We demonstrated the function of a novel multi-modal MR image synthesis neural network, GRMM-GAN, for T1ce MR synthesis based on pre-contrast T1 and T2 MR images. GRMM-GAN shows promise for avoiding repeated contrast injections during radiation therapy treatment.
ARTICLE | doi:10.20944/preprints202308.0131.v1
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: deep generative model (DGM); Variational Autoencoders (VAE); Generative Adversarial Network (GAN)
Online: 2 August 2023 (03:39:21 CEST)
Generative artificial intelligence (GenAI) has been developing with many incredible achievements like ChatGPT and Bard. Deep generative model (DGM) is a branch of GenAI, which is preeminent in generating raster data such as image and sound due to strong points of deep neural network (DNN) in inference and recognition. The built-in inference mechanism of DNN, which simulates and aims to synaptic plasticity of human neuron network, fosters generation ability of DGM which produces surprised results with support of statistical flexibility. Two popular approaches in DGM are Variational Autoencoders (VAE) and Generative Adversarial Network (GAN). Both VAE and GAN have their own strong points although they share and imply underline theory of statistics as well as incredible complex via hidden layers of DNN when DNN becomes effective encoding/decoding functions without concrete specifications. In this research, I try to unify VAE and GAN into a consistent and consolidated model called Adversarial Variational Autoencoders (AVA) in which VAE and GAN complement each other, for instance, VAE is good at generator by encoding data via excellent ideology of Kullback-Leibler divergence and GAN is a significantly important method to assess reliability of data which is realistic or fake. In other words, AVA aims to improve accuracy of generative models, besides AVA extends function of simple generative models. In methodology this research focuses on combination of applied mathematical concepts and skillful techniques of computer programming in order to implement and solve complicated problems as simply as possible.
ARTICLE | doi:10.20944/preprints202306.1350.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: mm-wave; deep space exploration; satellite communication; power amplifier; GaN; HEMT
Online: 20 June 2023 (02:19:32 CEST)
This paper presents a 32 GHz high power amplifier (HPA) with a design strategy to achieve a high-power output with reliable operation for the Ka-band deep space satellite communication in 150 nm GaN HEMT technology. The presented Ka-band HPA employs a cascaded two-stage common source amplifier topology, and the output stage is comprised of an 8-way power combining network in current mode. The interstage matching network is designed with the bandpass configuration utilizing capacitors and transmission lines to provide better stability at the low-frequency regime. The implemented Ka-band HPA achieved a peak power gain of 7.3 dB at 32 GHz and a 3 dB bandwidth was 3.5 GHz between 31.3 and 34.8 GHz. The saturated output power at the peak power added efficiency (PAE) of 19.3% was 38.2 dBm, and the output 1 dB gain compression point (OP1dB) was 27.4 dBm in the measurement. The designed HPA consumes an area of 19.35 mm^2 including RF signal, and DC pads
ARTICLE | doi:10.20944/preprints201907.0049.v1
Subject: Chemistry And Materials Science, Nanotechnology Keywords: core/shell nanowires; GaN; Ga2O3; metal-oxide semiconductor; gas sensor devices
Online: 3 July 2019 (08:33:38 CEST)
The development of sensors for monitoring Carbon Monoxide (CO) in a large range of temperature is of crucial importance in areas as monitoring of industrial processes or personal tracking using smart objects. Devices integrating GaN/Ga2O3 core/shell nanowires (NWs) is a promising solution allowing combining the high sensitivity of the electronic properties to the states of GaN-core surface; and the high sensitivity to CO of Ga2O3-shell. Because the performances of sensors primarily depend on the material properties composing the active layer of the device, it is essential to control these properties and in first time its synthesis. In this work, we investigate the synthesis of GaN/Ga2O3 core-shell NWs with a special focus on the formation of the shell. The GaN NWs grown by Plasma-assisted molecular beam epitaxy, are post-treated following thermal oxidation to form Ga2O3-shell surrounding the GaN-core. We establish that the Ga2O3-shell thickness can be modulated from 1 up to 14 nm by changing the oxidation conditions, and follows the diffuse-controlled reaction. By combining XRD-STEM and EDX analysis, we also demonstrate that the oxide shell formed by thermal oxidation is crystalline and presents the β-Ga2O3 crystalline phase, and is synthesized in epitaxial relationship with the GaN-core.
REVIEW | doi:10.20944/preprints202309.1820.v1
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: Breast Cancer; Deep Learning Methods; Image Classification; GAN; Transfer Learning; Lifelong Learning
Online: 27 September 2023 (05:17:09 CEST)
Breast cancer is a common malignant tumour and studies have shown that early and accurate detection is crucial for patients. With the maturity of medical imaging and deep learning development, significant progress has been made in breast cancer classification, which greatly improves the accuracy and efficiency of classification. This review focuses on deep learning, migration learning, GAN, and lifelong learning to elaborate and summarise the important roles arising from breast cancer detection. This review also examines the dataset and labeling issues required for breast cancer classification. In conclusion, at the end of the article, we look at future directions for breast cancer classification research, including cross-migration learning, multimodal data fusion, model interpretability, and lifelong learning, and also explore how to provide personalized treatment plans for patients.
ARTICLE | doi:10.20944/preprints202308.1958.v1
Subject: Chemistry And Materials Science, Materials Science And Technology Keywords: GaN nanowires; ZrN buffer layers; molecular beam epitaxy; nanowire orientation; geometrical selection
Online: 30 August 2023 (04:11:04 CEST)
GaN nanowires grown on metal substrates attract an increasing interest for a wide range of applications. Herein, we report GaN nanowires grown by plasma-assisted molecular beam epitaxy on thin polycrystalline ZrN buffer layers, sputtered onto Si(111) substrates. The nanowire orientation is studied experimentally and described within a model as a function of the Ga beam angle, nanowire tilt angle and substrate rotation. We show that vertically aligned nanowires grow faster than inclined nanowires, which leads to an interesting effect of geometrical selection of the nanowire orientation in the directional molecular beam epitaxy technique. After a given growth time, this effect depends on the nanowire surface density. At low density, the nanowires continue to grow with random orientations as nucleated. At high density, the effect of preferential growth induced by unidirectional supply of material in MBE starts to dominate. Faster growing nanowires with smaller tilt angles shadow more inclined nanowires that grow slower. This helps to obtain more regular ensembles of vertically oriented GaN nanowires despite their random position induced by the metallic grains at nucleation. The obtained dense ensembles of vertically aligned GaN nanowires on ZrN/Si(111) surfaces are highly relevant for device applications. These results are not specific for GaN nanowires and should be relevant for any nanowires which are epitaxially linked to the randomly oriented surface grains in the directional molecular beam epitaxy.
ARTICLE | doi:10.20944/preprints202307.0804.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: AlGaN/GaN HEMT device; CCM; DCM; dynamic on‐resistance; dynamic power loss
Online: 12 July 2023 (09:35:11 CEST)
In this work, we present an analytical model of dynamic power losses for enhancement-mode AlGaN/GaN high-electron-mobility transistor power devices (eGaN HEMTs). To build this new model, the dynamic on-resistance (Rdson) is first accurately extracted by our ex-traction circuit based on a double-diode isolation (DDI) method for a high operating frequency up to 1 MHz and a large drain voltage up to 600 V, thus the unique problem of an increase in the dynamic Rdson is presented. Then the impact of the current operation mode on the on/off transi-tion time is evaluated by a dual-pulse-current-mode test (DPCT) which including a discontinu-ous conduction mode (DCM) and a continuous conduction mode (CCM), thus the transition time is revised for different current mode. Afterward the discrepancy between the drain current and the real channel current is qualitative investigated by an external shunt capacitance (ESC) meth-od, thus the losses due to device parasitic capacitance are also taken into account. After these improvements, the dynamic model will be more compatible for eGaN HEMTs. Finally, the dynamic power losses calculated by this model are verified to be in good agreement with experimental results. Based on this model, we propose a superior solution with qua-si-resonant mode (QRM) to achieve lossless switching and accelerate switching speed.
ARTICLE | doi:10.20944/preprints202307.0666.v1
Subject: Engineering, Bioengineering Keywords: Compressed sensing MRI; GAN; U-net; dilated-residual blocks; channel attention mechanism
Online: 11 July 2023 (10:23:45 CEST)
Compressed Sensing (CS) MRI has shown great potential in enhancing time efficiency. Deep learning techniques, specifically Generative Adversarial Networks (GANs), have emerged as potent tools for speedy CS-MRI reconstruction. Yet, as the complexity of deep learning recon-struction models increases, this can lead to prolonged reconstruction time and challenges in achieving convergence. In this study we present a novel GAN-based model that delivers superior performance without escalating model complexity. Our generator module, built on the U-net architecture, incorporates dilated residual (DR) networks, thus expanding the network's receptive field without increasing parameters or computational load. At every step of the downsampling path, this revamped generator module includes a DR network, with the dilation rates adjusted according to the depth of the network layer. Moreover, we have introduced a channel attention mechanism (CAM) to distinguish between channels and reduce background noise, thereby fo-cusing on key information. This mechanism adeptly combines global maximum and average pooling approaches to refine channel attention. We conducted comprehensive experiments with the designed model using public domain MRI datasets of the human brain. Ablation studies af-firmed the efficacy of the modified modules within the network. Compared to other relevant models, our proposed model exhibits exceptional performance, achieving not only excellent sta-bility but also outperforming other networks in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). The model presents a promising pathway for enhancing the efficiency and quality of CS-MRI reconstruction.
ARTICLE | doi:10.20944/preprints202311.0161.v1
Subject: Computer Science And Mathematics, Computer Vision And Graphics Keywords: deep learning; skin cancer; image augmentation; GAN; geometric augmentation; image classification; interpretable technique
Online: 2 November 2023 (10:52:57 CET)
This research paper presents a deep learning approach to early detection of skin cancer using image augmentation techniques. The authors propose a two-stage image augmentation technique that involves the use of geometric augmentation and generative adversarial network (GAN) to classify skin lesions as either benign or malignant. This research utilized the public HAM10000 dataset to test the proposed model. Several pre-trained models of CNN were employed, namely Xception, Inceptionv3, Resnet152v2, EfficientnetB7, InceptionresnetV2, and VGG19. Our approach achieved accuracy, precision, recall, and F1-score of 96.90%, 97.07%, 96.87%, 96.97%, respectively, which is higher than the performance achieved by other state-of-the-art methods. The paper also discusses the use of SHapley Additive exPlanations (SHAP), an interpretable technique for skin cancer diagnosis, which can help clinicians understand the reasoning behind the diagnosis and improve trust in the system. Overall, the proposed method presents a promising approach to automated skin cancer detection that could improve patient outcomes and reduce healthcare costs.
ARTICLE | doi:10.20944/preprints202304.0086.v1
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: Image fusion; generative adversarial network (GAN); local binary patterns (LBP); multi-modal images
Online: 6 April 2023 (10:03:31 CEST)
Image fusion is the process of combining multiple input images from single or multiple imaging modalities into a fused image, which is expected to be more informative for human or machine perception as compared to any of the input images. In this paper, we propose a novel method based on deep learning for fusing infrared images and visible images, named the LBP-based proportional input generative adversarial network (LPGAN). In the image fusion task, the preservation of structural similarity and image gradient information is contradictory, and it is difficult for both to achieve good performance at the same time. To solve this problem, we innovatively introduce Local Binary Patterns (LBP) into Generative Adversarial Networks (GANs), which effectively utilize the texture features of the source images, so that the network has stronger feature extraction ability and anti-interference ability. In the feature extraction stage, we introduce a pseudo-siamese network for the generator to extract the detailed features and the contrast features. At the same time, considering the characteristic distribution of different modal images, we propose a 1:4 scale input mode. Extensive experiments on the publicly available TNO dataset and CVC14 dataset show that the proposed method achieves the state-of-the-art performance. We also test the universality of LPGAN through the fusion of RGB and infrared images on the RoadScene dataset. In addition, LPGAN is applied to multi-spectral remote sensing image fusion. Both qualitative and quantitative experiments demonstrate that our LPGAN can not only achieve good structural similarity, but also retain rich detailed information.
ARTICLE | doi:10.20944/preprints202006.0056.v1
Subject: Chemistry And Materials Science, Metals, Alloys And Metallurgy Keywords: Microstructure Modelling; Representative Volume Elements; DP-steel; Machine Learning; Deep Learning; Wasserstein GAN
Online: 5 June 2020 (14:37:20 CEST)
For the generation of representative volume elements a statistical description of the relevant parameters is necessary. These parameters usually describe the geometric structure of a single grain. Commonly, parameters like area, aspect ratio and slope of the grain relative to the rolling direction are applied. However, usually simple distribution functions like log normal or gamma distribution are used. Yet, these do not take the interdependencies between the microstructural parameters into account. To fully describe any metallic microstructure though, these interdependencies between the singular parameters need to be accounted for. To accomplish this representation, a machine learning approach was applied in this study. By implementing a Wasserstein generative adversarial network, the distribution, as well as the interdependencies could accurately be described. A validation scheme was applied to verify the excellent match between microstructure input data and synthetically generated output data.
Subject: Chemistry And Materials Science, Applied Chemistry Keywords: Diluted magnetic semiconductors (DMS), Manganese Mn-doped GaN (Mn.Gac.N), Spintronic, Opto-electronic devices
Online: 17 May 2019 (16:16:22 CEST)
Sensors became integrated through the control condition arrangement, either for visual, mechanical, biological, or chemical applications. New stuff is designed for detection, as Diluted Magnetic Semiconductors (DMS), and are considered attractive candidates that consist of a traditional 111- V, II-VI, or group IV semiconductor. Manganese Mn-doped GaN (Mn.Gac.N) epitaxial velum has a unique magnetic, visual and chemical properties in order to control system intelligent for detector design. The subject area of the magnetic properties of MnxGal-xN is on a large scale available, there are only a few studies on the visual properties and electrochemical properties of MnxGal-xN epitaxial velums. Where MnGaN velums were used in spintronic and opto-electronic applications according to their magnetic characterization and constructed MnGaN electrode are drop fabric potential for potentiometric sensor applications since they have good performance as ion-selective electrodes. The electrical and magnetic properties that allow the control of electron spin as well as complaint period, makes the materials ideal for spintronic applications. Designing such spintronic and optoelectronic devices based on MnxGal-xN requires a broader agreement of physical, visual, electrical and chemical properties epitaxial velums that are still seldom in the literature. This bailiwick displays the potential use of MnGaN semiconductor as an all solid-state potentiometric sensor for measuring anions in solutions in the control engineering field.
REVIEW | doi:10.20944/preprints201812.0072.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: gallium nitride (GaN); silicon carbide (SiC); silicon (Si); converters; power devices; power electronics
Online: 6 December 2018 (05:16:44 CET)
Power semiconductor devices are essential from the operation point of view, size, efficiency and cost, these components are used in a myriad of applications, providing features that make them an important part of the system in which they are operating. This document analyzes and compares the basic structure, properties, design aspects, as well as temperature performance, stability and switching losses, present in devices on silicon (Si), silicon carbide (SiC) and new generation devices fabricated in gallium nitride (GaN) applied in renewable energy systems. The main objective is determinate the viability of the new generation components, which present a superior performance in view of an increase in efficiency, conductivity, decreases in switching losses, lower resistances and parasitic capacitances as well as higher operating frequency range. Therefore demonstrating the GaN components are a strong and viable candidate to solve some of the problems present in renewable energy systems.
ARTICLE | doi:10.20944/preprints202310.0073.v1
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: quantum computing; quantum error correction; quantum sampling; quantum information theory; GAN; pix2pix; LPIQE; PDU
Online: 3 October 2023 (03:09:23 CEST)
Quantum image representation is a widely researched area of quantum computing. Currently developed methods use angle parameter of the rotation gate (e.g., the FRQI method), sequences of qubits (ex. NEQR method) or phase shift (ex. LPIQE method) for storing color information of pixels. All of those methods are affected by decoherence and other classical and quantum noise, which is an inseparable part of quantum computing in a NISQ (Noisy Intermediate Scale Quantum) era. These all phenomenons influence the measurements, its probability distribution over the measurement basis and, as the result, makes the extracted images not even similar to those, which was stored in quantum computers. Since this process is, in its foundation, quantum as well, the computational reversal of this process is possible. There are a lot of methods for error correction, mitigation and reduction, but all of them use quantum computer time, or additional qubits to achieve desired result. We report a successful use of the Generative Adversarial Network tuned for image-to-image translation in conjunction with PDU method, for error reduction in images encoded using LIPQE (Local Phase Image Quantum Encoding) method.
COMMUNICATION | doi:10.20944/preprints202305.1735.v1
Subject: Chemistry And Materials Science, Materials Science And Technology Keywords: InGaN/GaN heterostructures; transmission electron microscopy; indium composition; screw-type dislocations; edge-type misfit dislocations
Online: 25 May 2023 (04:37:06 CEST)
We have investigated the interfacial dislocations in InxGa1–xN/GaN (0 ≤ x ≤ 0.20) heterostructures using diffraction contrast analysis in a transmission electron microscopy. The analysis indicate that the structural properties of the interface dislocations depend on the indium composition. For lower indium composition up to x = 0.09, we observed that the screw-type dislocations and dislocation half loops occurred at the interface even though the former do not contributes toward elastic relaxation of the misfit strain in the InGaN layer. With the increase of indium composition (0.13 ≤ x ≤ 0.17), in addition to the network of screw-type dislocations, edge-type misfit dislocations were found generated with their density gradually increasing. For higher indium composition (0.18 ≤ x ≤ 0.20), all the interfacial dislocations are transformed into a network of straight misfit dislocations along the <10-10> directions leading to partially relaxation of the InGaN epilayer. The presence of dislocation half loops may be explained by slip on the basal plane, the formation of misfit dislocations are attributed to punch-out mechanism.
ARTICLE | doi:10.20944/preprints202011.0039.v1
Subject: Computer Science And Mathematics, Computer Vision And Graphics Keywords: generative model; human movement; conditional Deep Convolutional Generative Adversarial Network; GAN; spatio-temporal pseudo-image
Online: 2 November 2020 (12:55:22 CET)
Generative models for images, audio, text and other low-dimension data have achieved great success in recent years. Generating artificial human movements can also be useful for many applications, including improvement of data augmentation methods for human gesture recognition. The object of this research is to develop a generative model for skeletal human movement, allowing to control the action type of generated motion while keeping the authenticity of the result and the natural style variability of gesture execution. We propose to use a conditional Deep Convolutional Generative Adversarial Network (DC-GAN) applied to pseudo-images representing skeletal pose sequences using Tree Structure Skeleton Image format. We evaluate our approach on the 3D-skeleton data provided in the large NTU RGB+D public dataset. Our generative model can output qualitatively correct skeletal human movements for any of its 60 action classes. We also quantitatively evaluate the performance of our model by computing Frechet Inception Distances, which shows strong correlation to human judgement. Up to our knowledge, our work is the first successful class-conditioned generative model for human skeletal motions based on pseudo-image representation of skeletal pose sequences.
ARTICLE | doi:10.20944/preprints201809.0498.v1
Subject: Medicine And Pharmacology, Oncology And Oncogenics Keywords: nasopharyngeal carcinoma; traditional Chinese medicine; Chinese herbal products; complementary and alternative medicine; Gan-Lu-Yin
Online: 26 September 2018 (05:09:08 CEST)
In most countries, the incidence of nasopharyngeal carcinoma (NPC) is no more than 1 per 100,000 for both men and women; however, it is much higher for men and women in Taiwan. The use of traditional Chinese medicine (TCM) as complementary and alternative medicine for the treatment of NPC and its treatment-related side effects has been increasing. The National Health Insurance (NHI) covers 99.6% of Taiwan’s residents. In the present population-based cohort study, we aimed to investigate the pattern of utilization of Chinese herbal products (CHPs) for NPC from 2001 through 2011 in Taiwan. We identified a total of 30294 patients with newly diagnosed NPC from the Registry for Catastrophic Illnesses Patient Database (RCIPD). Descriptive statistics and multiple logistic regression analysis were employed to estimate the adjusted odds ratios (aORs) for CHP utilization. From 2001 through 2011, 17816 patients aged ≥20 years were newly diagnosed with NPC. Of these, 4749 patients used TCM outpatient services for NPC treatment. TCM users were more likely to be women, young, residents of Central Taiwan, and white-collar workers. The most commonly prescribed formula CHP was Gan-Lu-Yin, followed by Xin-Yi-Qing Fei-Tang and Shan-Shen-Mai-Men-Dong-Tang. The most commonly prescribed single CHP was Hedyotis diffusa, followed by Radix Scrophulariae and Radix Ophiopogonis. These findings provide information regarding personalized therapies for NPC and can promote further clinical experiments and pharmacological research on CHPs for NPC treatment in Taiwan. Further well-designed randomized controlled studies and basic mechanistic studies should assess the safety and effectiveness of CHPs for NPC treatment.
ARTICLE | doi:10.20944/preprints201609.0125.v2
Subject: Chemistry And Materials Science, Other Keywords: GaN ultraviolet photodetector; periodic trapezoid column-shape patterned sapphire substrate; responsivity; UV-to-visible rejection ratio
Online: 18 October 2016 (08:19:48 CEST)
GaN ultraviolet photodetector with metal-semiconductor-metal structure is achieved by growing on a periodic trapezoid column-shape patterned sapphire substrate using metalorganic chemical vapor deposition. Under 5-V reverse bias, the photodetector fabricated on such patterned sapphire substrate exhibits a lower dark current, a higher photocurrent, and a 476 % enhancement in the maximum responsivity as compare with those of the photodetector fabricated on conventional flat sapphire substrate. It is also found that the much larger UV-to-visible rejection ratio and the fact that responsivity drops in a smaller cut-off region are observed from photodetector fabricated by using a periodic trapezoid column-shape patterned sapphire substrate. These phenomena may all be attributed to the reduction of threading dislocation density and the improved quality of GaN film, as well as the internal reflection and/or scattering effect on the interface between GaN film and the periodic trapezoid column-shape pattern of the substrate.
Subject: Computer Science And Mathematics, Artificial Intelligence And Machine Learning Keywords: 2019 novel coronavirus; COVID-19; SARS-CoV-2; Deep Transfer Learning; Convolutional Neural Network; Machine Learning; GAN
Online: 7 April 2020 (10:59:04 CEST)
The coronavirus (covid-19) pandemic is putting healthcare systems across the world under unprecedented and increasing pressure according to the World Health Organization (WHO). With the advances in computer algorithms and especially Artificial Intelligence, the detection of this type of virus in the early stages will help in fast recovery and help in releasing the pressure off healthcare systems. In this paper, a GAN with deep transfer learning for coronavirus detection in chest x-ray images is presented. The lack of benchmark datasets for covid-19 especially in chest x-rays images is the main motivation of this research. The main idea is to collect all the possible images for covid-19 that exists until the writing of this research and use the GAN network to generate more images to help in the detection of the virus from the available x-rays images with the highest accuracy possible. The dataset used in this research was collected from different sources and it is available for researchers to download and use it. The number of images in the collected dataset is 307 images for four different types of classes. The classes are the covid-19, normal, pneumonia bacterial, and pneumonia virus. The dataset is divided into 90% for the GAN and the training and the validation phase, while 10% used in the testing phase. The GAN helps in generating more images from the original dataset to be 30 times larger than the originally collected dataset. The GAN also help in overcoming the overfitting problem and made the proposed model more robust. Three deep transfer models are selected in this research for investigation. The models are the Alexnet, Googlenet, and Restnet18. Those models are selected based on their small number of layers on their architectures, which will reflect in reducing the complexity of the models and the consumed memory and time. Using a combination of GAN and deep transfer models prove it is efficiency according to validation, testing accuracy, and performance measurements such as precision, recall, and F1 score. Three case scenarios are tested through the paper, the first scenario which includes 4 classes from the dataset, while the second scenario includes 3 classes and the third scenario includes 2 classes. All the scenarios include the covid-19 class as it is the main target of this research to be detected. In the first scenario, the Googlenet is selected to be the main deep transfer model as it achieves 80.6% in testing accuracy. In the second scenario, the Alexnet is selected to be the main deep transfer model as it achieves 85.2% in testing accuracy, while in the third scenario which includes 2 classes(covid-19, and normal), Googlenet is selected to be the main deep transfer model as it achieves 100% in testing accuracy and 99.9% in the validation accuracy. All the performance measurement strengthen the obtained results through the research. Finally, this research may be considered one of the first trails to use GAN and deep transfer models together to help in detecting coronaviruses (covid-19) within the absence of a benchmark dataset around the world, especially in x-rays chest images.
ARTICLE | doi:10.20944/preprints202010.0502.v1
Subject: Environmental And Earth Sciences, Atmospheric Science And Meteorology Keywords: Statistical downscaling; Generative Adversarial Network; Combination of Errors; Convolutional Neural Network; multi-scale structural similarity index; Wasserstein GAN
Online: 25 October 2020 (19:33:49 CET)
Despite numerous studies in statistical downscaling methodology, there remains a lack of methods that can downscale from precipitation modeled in global climate models to regional level high resolution gridded precipitation. This paper reports a novel downscaling method using a Generative Adversarial Network (GAN), CliGAN, which can downscale large-scale annual maximum precipitation given by simulation of multiple atmosphere-ocean global climate models (AOGCM) from Coupled Model Inter-comparison Project 6 (CMIP6) to regional-level gridded annual maximum precipitation data. This framework utilizes a convolution encoder-dense decoder network to create a generative network and a similar network to create a critic network. The model is trained using an adversarial training approach. The critic uses the Wasserstein distance loss function and the generator is trained using a combination of adversarial loss Wasserstein distance, structural loss with the multi-scale structural similarity index (MSSIM), and content loss with the Nash-Sutcliff Model Efficiency (NS). The MSSIM index allowed us to gain insight into the model’s regional characteristics and shows that relying exclusively on point-based error functions, widely used in statistical downscaling, may not be enough to reliably simulate regional precipitation characteristics. Further use of structural loss functions within CNN-based downscaling methods may lead to higher quality downscaled climate model products.
ARTICLE | doi:10.20944/preprints202205.0026.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: analog circuit design; buffer amplifiers; offset voltage's systematic component; voltage followers; operational amplifiers; depletion-mode; СMOS; JFET; Si; GaAs; GaN
Online: 5 May 2022 (08:42:15 CEST)
The authors of the article performed computer simulation of buffer amplifiers (BA), which have medium and extremely small values of the offset voltage's systematic component (Voff), for different technological processes (Si, GaAs and GaN). The proposed control units are distinguished by a small number of elements and allow operation in the range of low and high temperatures. The variants of circuitry implementation of control units based on GaAs, GaN depletion-mode CMOS and JFET technological processes are considered. The results of the comparative modeling showed that the basic circuit of the BA on two field-effect transistors, when implemented on various modifications of GaN MOS and depletion-mode MOS transistors, provides sufficiently low values of the offset voltage's systematic component (less than 2 μV). The proposed BAs are designed for use in the structure of the Sallen-Key low-pass filter (LPF) when they are implemented both on mid-frequency Si CJFET and on GaAs microwave transistors. Low values of the LPF Voff have a positive effect on the effective capacity of the ADC. An example of switching on a BA in the JFET OpAmp structure based on the depletion-mode MOS input stage and a “folded” cascode, which, with 100% negative feedback, can be used in the Sallen-Key LPF, is considered. Computer simulation of the JFET/MOS OpAmp showed that the OpAmp has an open-loop voltage gain of 76-85dB, and its Voff is within 7µV in the temperature range from -60°C to +120°C. The presented circuitry of buffer amplifiers is intended, first of all, for the tasks of designing precision Sallen-Key low-pass filter (low-pass filter, high-pass filter, PF, RF).
ARTICLE | doi:10.20944/preprints201811.0412.v1
Subject: Engineering, Control And Systems Engineering Keywords: Geographical Area Network (GAN); Structural Health Monitoring (SHM); Utility Computing (UC); Things as a Service (TaaS); Internet of Things (IoT)
Online: 19 November 2018 (03:58:56 CET)
In view of intensified disasters and fatalities caused by natural phenomena and geographical expansion, there is a pressing need for a more effective environment logging for a better management and urban planning. This paper proposes a novel utility computing model (UCM) for structural health monitoring (SHM) that would enable dynamic planning of monitoring systems in an efficient and cost-effective manner. The proposed UCM consists of network-attached data drive that stores data from SHM logger, population count system and Geographic Information System (GIS) enhanced with a Cloud IoT data backup, display, and analysis server. The UCM using this data and data from building information systems applies a simple machine learning algorithm to generate real-time structure health and suggests re-planning of SHM units. The health of structure varies dynamically with disturbances created by higher occupancy and structure density per zone. The proposed SHM-UCM is unique in terms of its capability to manage heterogeneous SHM resources. This was tested in a case study on Qatar University (QU) in Doha Qatar, where it looked at where SHM nodes are distributed along with occupancy density in each building. This information was taken from QU simulated occupation and zone calculation models and then compared to ideal SHM system data. Results show the effectiveness of the proposed model in logging and dynamically planning SHM.
ARTICLE | doi:10.20944/preprints202207.0356.v1
Subject: Environmental And Earth Sciences, Atmospheric Science And Meteorology Keywords: Multi-site statistical downscaling; Generative Adversarial Network; Combination of Errors; Convolutional Neural Network; Struc-tural Similarity Index; Wasserstein GAN; extreme precipitation
Online: 25 July 2022 (07:59:59 CEST)
Although the statistical methods of downscaling climate data have progressed significantly, the development of high-resolution precipitation products continues to be a challenge. This is especially true when interest centres on downscaling value over several study sites. In this paper , we report a new downscaling method termed the multi-site Climate Generative Adversarial Network (MSCliGAN), which can simulate annual maximum precipitation to the regional scale during the 1950-2010 period in different cities in Canada by using different AOGCM's from the Coupled Model Inter-Comparison Project 6 (CMIP6) as input. Auxiliary information provided to the downscaling model included topography and land-cover. The downscaling framework uses a convolution encoder-decoder U-net network to create a generative network and a convolution encoder network to create a critic network. An adversarial training strategy is used to train the model. The critic/discriminator used Wasserstein distance as a loss measure and on the other hand the generator is optimized using a summation of content loss on Nash-Shutcliff Model Efficiency (NS), structural loss on structural similarity index (SSIM), and adversarial loss Wasserstein distance. Downscaling results show that downscaling AOGCMs by incorporating topography and land-use/land-cover can produce spatially coherent fields close to observation over multiple-sites. We believe the model has sufficient downscaling potential in data sparse regions where climate change information is often urgently needed.
ARTICLE | doi:10.20944/preprints202010.0010.v1
Subject: Engineering, Electrical And Electronic Engineering Keywords: Totem-pole power factor correction; energy storage systems (ESS); digital control; Gallium Nitride (GaN) based; current harmonic distortion mitigation; efficiency and power quality improvement
Online: 1 October 2020 (09:12:36 CEST)
With the unceasing advancement in wide-bandgap (WBG) semiconductor technology, the minimal reverse-recovery charge Qrr and other more powerful natures of WBG transistors enable totem-pole bridgeless PFC to become a dominant solution for energy storage systems (ESS). This paper focuses on design and implementation of a control structure for a totem-pole boost PFC with newfangled enhancement-mode Gallium Nitride (eGaN) FETs, not only to simplify the control implementation, but also to achieve high power quality and efficiency. The converter is designed to convert a 90-264-VAC input to a 385-VDC output for a 2.6-kW output power. Lastly, to validate the methodology, an experimental prototype is characterized and fabricated. The uttermost efficiency at 230 VAC attains 99.14%. The lowest total harmonic distortion in the current (ITHD) at high line condition (230 V) reaches 1.52% while the power factor gains 0.9985.