Microstructure Generation via Generative Adversarial Network for Heterogeneous, Topologically Complex 3D Materials

Polarization resistance of solid oxide fuel cell electrodes consists of resistances due to mass transfer and chemical reaction. Therefore, higher ion conductivity, electron conductivity, and gas diffusivity and larger triple-phase boundary (TPB) density lead to lower polarization resistance. To increase effective conductivities and diffusivity, tortuosity factors in each mass-transfer path should be reduced. In this study, magnetic alignment method is developed to align the mass-transfer paths in the electrode microstructure. Since Ni as an electron-conducting material in anode has high magnetic permeability, the particles are polarized and aligned by a magnetic field. Therefore, the tortuosity factors can be reduced. In addition, this Ni alignment is expected to reduce the tortuosity factors of yttria-stabilized zirconia (YSZ) and pore phases as well. However, there is a risk for a decrease in TPB density. The effect of magnetic field on resultant microstructure and polarization resistance of Ni-YSZ anode are investigated by impedance measurement of anode symmetrical cells and by reconstruction of three-dimensional microstructure. The tortuosity factors of Ni and YSZ decreased and TPB density decreased by the magnetic field. As a result, the polarization resistance of anode was kept nearly unchanged. The effect of reducing Ni volume fraction is also investigated.

Quantitative estimation of triple phase boundaries in solid oxide fuel cell electrodes via artificial neural network

Both JPEG restoration and anti-forensics remove the artifacts left by JPEG compression, and recover the JPEG compressed image. However, how to restore the high-frequency details of a JPEG compressed image for JPEG restoration and how to deceive the existing JPEG compression detectors without sacrificing visual quality in JPEG anti-forensics remain challenging. To address these issues, a framework of generative adversarial networks (GAN) with novel loss functions for JPEG restoration and anti-forensics (JRA-GAN) is proposed to allow a JPEG compressed image to be translated into a reconstructed one. Since JPEG compression causes impairment to high-frequency components, an alternating current (AC)-component loss function that measures the loss of AC components is proposed in JRA-GAN to recover these components. To prevent forensic detection, a calibration loss function is also introduced in JRA-GAN to mitigate the variance gap in the high-frequency subbands between a generated image and its calibrated version. Our experimental results demonstrate that the proposed JPEG restoration method outperforms existing methods in terms of image quality, and the JPEG anti-forensic scheme achieves better visual quality and anti-forensic performance that is comparable to the existing state-of-the-art anti-forensic methods. Our code is available in this page: https://github.com/wujianyuan/JRG-GAN .

Design of optimal microstructures for infiltrated solid oxide fuel cell (SOFC) electrodes is a complicated process because of the multitude of the electrochemical and physical phenomena taking place in the electrodes in different temperatures, current densities and reactant flow rates. In this study, a stochastic geometric modeling method is used to create a range of digitally realized infiltrated SOFC electrode microstructures to extract their geometry-related electrochemical and physical properties. Triple Phase Boundary (TPB), active surface density of particles along with the gas transport factor is evaluated in those realized models to adapt for various infiltration strategies. Recently, additive manufacturing or freeze type casting methods enable researchers to investigate the performance of directional electrodes to get the maximum electrochemical reaction sites, gas diffusivity and ionic conductivity simultaneously. A series of directional backbones with different amount of virtually deposited electrocatalyst particles are characterized in the first step. The database of microstructural parameters (inputs) and effective geometric properties (outputs) is used to train a range neural network. A microstructure property hull is created using the best neural network model to discover the range of effective properties, their relative behaviour and optimum microstructure. The characteristics of models is shown that there is not any contradiction between the high level of TPB and contact surface density of particles, but the highest amount of gas diffusivity can be found in the microstructures with lower level of reaction sites. Also increasing the contact surface density has a negative effect on gas transport but the high level of TPB density is feasible in the full range of microstructures. In the other hand, TPB density and gas diffusion into the models are inversely related, although there are a limited number of microstructures with high level of reaction sites and acceptable gas diffusivity. Finally, using a simple optimization process, the microstructures with the highest level of reaction sites and gas transport factor are identified which have the backbone porosity of about 50%, and extremely higher gain growth rate normal to the electrolyte. Additive manufacturing and 3D printing methods will enhance researchers in the future to create the real directional electrodes on the base of these proposed models.

Investigation of the active thickness of solid oxide fuel cell electrodes using a 3D microstructure model

Generative Adversarial Networks (GANs) have received a great deal of attention due in part to recent success in generating original, high-quality samples from visual domains. However, most current methods only allow for users to guide this image generation process through limited interactions. In this work we develop a novel GAN framework that allows humans to be “in-the-loop” of the image generation process. Our technique iteratively accepts relative constraints of the form “Generate an image more like image A than image B”. After each constraint is given, the user is presented with new outputs from the GAN, informing the next round of feedback. This feedback is used to constrain the output of the GAN with respect to an underlying semantic space that can be designed to model a variety of different notions of similarity (e.g. classes, attributes, object relationships, color, etc.). In our experiments, we show that our GAN framework is able to generate images that are of comparable quality to equivalent unsupervised GANs while satisfying a large number of the constraints provided by users, effectively changing a GAN into one that allows users interactive control over image generation without sacrificing image quality.

As a new unsupervised learning algorithm framework, generative adversarial networks(GAN) have been favored by more and more researchers, and it has become a research hotspot now. GAN is inspired by the two-person zero-sum game theory in game theory. Its unique adversarial training idea can generate high-quality samples and has more powerful feature learning and feature expression capabilities than traditional machine learning algorithms. At present, GAN has achieved remarkable success in the field of computer vision, especially in the field of sample generation. Every year, a large number of GAN-related research papers are produced, reflecting the fiery degree of research on GAN model. Aiming at the hot model of GAN, first introduce the research status of GAN; then introduce the theory and framework of GAN, which analyzes in detail why the gradient disappears and the mode collapses during the training of GAN; then discussed some typical GAN improvement models, and summarized their theoretical improvements, advantages, limitations, application scenarios and implementation costs; Finally, the application results of GAN in data generation, image super-resolution, and image style conversion are shown, and the current challenges and future research directions of GAN are discussed.

In a deep-learning-based algorithm, generative adversarial networks can generate images similar to an input. Using this algorithm, an artificial three-dimensional (3D) microstructure can be reproduced from two-dimensional images. Although the generated 3D microstructure has a similar appearance, its reproducibility should be examined for practical applications. This study used an automated serial sectioning technique to compare the 3D microstructures of two dual-phase steels generated from three orthogonal surface images with their corresponding observed 3D microstructures. The mechanical behaviors were examined using the finite element analysis method for the representative volume element, in which finite element models of microstructures were directly constructed from the 3D voxel data using a voxel coarsening approach. The macroscopic material responses of the generated microstructures captured the anisotropy caused by the microscopic morphology. However, these responses did not quantitatively align with those of the observed microstructures owing to inaccuracies in reproducing the volume fraction of the ferrite/martensite phase. Additionally, the generation algorithm struggled to replicate the microscopic morphology, particularly in cases with a low volume fraction of the martensite phase where the martensite connectivity was not discernible from the input images. The results demonstrate the limitations of the generation algorithm and the necessity for 3D observations.

Effect of powder morphology on the microstructural characteristics of La0.6Sr0.4Co0.2Fe0.8O3 cathode: A Kinetic Monte Carlo investigation

This paper presents MCI-GAN, a novel menstrual cycle imputation (MCI) and generative adversarial network (GAN) framework designed to address the challenge of missing pixel imputation in medical images. Inspired by the intelligent behavior of the endometrial lining during the menstrual cycle, our method introduces four key innovations. First, we propose a novel metaheuristic algorithm that assigns weights to surround pixels based on menstrual cycle behavior, ensuring that the imputed pixels maintain structural integrity and coherence with their neighbors, thus preserving overall image quality. Second, to enhance the learning capability of the GAN, identity blocks are integrated into the network architecture, improving the network’s ability to capture complex spatial relationships and leading to more accurate and consistent imputation of missing pixels. Third, we introduce an adaptive loss function that dynamically adjusts the penalty for pixel discrepancies based on local image context, allowing the model to focus on areas where accurate imputation is most critical and thereby enhancing overall image fidelity. Fourth, the framework incorporates a multi-scale feature extraction mechanism, enabling the GAN to process and combine information at various levels of detail, ensuring that both fine-grained textures and larger structural patterns are accurately captured during the imputation process. The efficacy of MCI-GAN is demonstrated across three diverse medical imaging datasets: mammograms, magnetic resonance imaging (MRI) scans, and skin lesion images. Our results show that the proposed method significantly outperforms existing approaches in terms of accuracy and structural coherence, offering a robust solution for missing pixel imputation in medical imaging.

Mixture model (MM) is a probabilistic framework allows us to define dataset containing $K$ different modes. When each of the modes is associated with a Gaussian distribution, we refer to it as Gaussian MM or GMM. Given a data point x, a GMM may assume the existence of a random index $k$ ∊ {1, …, K} identifying which Gaussian the particular data is associated with. In a traditional GMM paradigm, it is straightforward to compute in closed-form, the conditional likelihood p(x|k, θ) as well as the responsibility probability p(k| x, θ) describing the distribution weights for each data. Computing the responsibility allows us to retrieve many important statistics of the overall dataset, including the weights of each of the modes/clusters. Modern large datasets are often containing multiple unlabelled modes, such as paintings dataset may contain several styles; fashion images containing several unlabelled categories. In its raw representation, the Euclidean distances between the data (e.g., images) do not allow them to form mixtures naturally, nor it's feasible to compute responsibility distribution analytically, making GMM unable to apply. In this paper, we utilize the generative adversarial network (GAN) framework to achieve a plausible alternative method to compute these probabilities. The key insight is that we compute them at the data's latent space $z$ instead of x. However, this process of $z$ → $x$ is irreversible under GAN which renders the computation of responsibility p(k|x, θ) infeasible. Our paper proposed a novel method to solve it by using a socalled posterior consistency module (PCM). PCM acts like a GAN, except its generator $C$ PCM does not output the data, but instead it outputs a distribution to approximate p(k|x, θ). The entire network is trained in an “end-to-end” fashion. Trough these techniques, it allows us to model the dataset of very complex structure using GMM and subsequently to discover interesting properties of an unsupervised dataset, including its segments, as well as generating new “out-distribution” data by smooth linear interpolation across any combinations of the modes in a completely unsupervised manner.

Solid Oxide Fuel Cells (SOFCs) are widely considered to be an ideal power source across a range of industries in the near future due to their high efficiency, high power density, and fuel flexibility. However, for widespread usage of SOFCs to be feasible, certain disadvantages, such as their high operating temperature, electrochemical losses, and high cost of manufacturing, must be solved. In particular, the field of materials research occupies a unique position of being able to quickly and simultaneously solve many of these issues through the development of material compositions and structures that do not currently exist in the industry. Especially relevant to solid oxide fuel cells in particular, is the impact that the electrodes’ geometries and compositions have on the operating performance. The electrodes’ ability to quickly move reactants to the interface of the electrolyte and remove products to the bulk flow comes into play as the fuel cell operates at higher current densities and concentration losses become prominent. Therefore, the microstructures present throughout the electrodes’ depth plays a vital role in improving overall cell performance. However, current fuel cell research has not yet effectively captured and defined the connections of the microstructure parameters and measured their impact on cell behavior. Additionally, homogeneous electrode design and implementation have dominated the research and commercial space thus far. The technique of functional material grading, which has begun to see wide research and commercial usage for such problems as acoustic impedance matching, thermal control, and mechanical design will be leveraged for the enhancement of SOFC electrodes. Functionally graded electrodes have been used in solid oxide fuel cell research in recent years in an effort to improve the cell performance by altering the microstructure including porosity, particle size, and composition of electronic and ionic conductors near the triple phase boundary region. Normally the functional grading process adds additional fabrication steps making it less desirable as an optimization process and easy manufacturing method. However, by using an additive manufacturing process it eliminates the need for multiple discrete graded layers to be deposited to obtain a graded electrode functional layer. This study seeks to explore the correlations of geometry and structure parameters from the meso-scale through to the micro-scale level, together with mass transfer, ionic and electronic transport, and gas-surface electrochemical reactions inside the electrodes to guide future additive manufacturing of SOFCs. Outlined is the development of an effective medium model for simulating the performance of fuel cell microstructures as a function of specific operating ranges in temperature, pressure, and in fuels used. Smooth and continuous linear and nonlinear functional gradation profiles of porosity and material composition are studied within the framework of an effective medium boundary value problem where the electrochemical relations, such as the Butler-Volmer equation, percolating conduction paths, and gas diffusion are solved using the finite difference method. The form of these relations for homogeneous electrodes was analyzed by Costamagna, et. al [1], who also defined outstanding problems with regards to optimizing SOFC electrodes [2]. Globally optimal linear and nonlinear gradation profiles and cell parameters are derived as a function of particle sizes, ratio of conductivities, and desired operating conditions and scored based on their reduction of cell overpotentials.[1] Costamagna, P., P. Costa, and V. Antonucci, Micro-modelling of solid oxide fuel cell electrodes. Electrochimica Acta, 1998. 43(3-4), pp.375-394.[2] Costamagna, P., P. Costa, and E. Arato, Some more considerations on the optimization of cermet solid oxide fuel cell electrodes. Electrochimica Acta, 1998. 43(8), pp.967-972. Figure 1

Using the spatial structure of various indoor environments as prior knowledge, the robot would construct the map more efficiently. Autonomous mobile robots generally apply simultaneous localization and mapping (SLAM) methods to understand the reachable area in newly visited environments. However, conventional mapping approaches are limited by only considering sensor observation and control signals to estimate the current environment map. This paper proposes a novel SLAM method, map completion network-based SLAM (MCN-SLAM), based on a probabilistic generative model incorporating deep neural networks for map completion. These map completion networks are primarily trained in the framework of generative adversarial networks (GANs) to extract the global structure of large amounts of existing map data. We show in experiments that the proposed method can estimate the environment map 1.3 times better than the previous SLAM methods in the situation of partial observation.

Hashing, a useful solution for approximate nearest neighbor (ANN) search, is popular for large-scale image retrieval. In this paper, we presents a deep supervised hashing model for remote sensing image retrieval (RSIR) in the framework of generative adversarial networks (GAN), named GAN-assist Hashing (GAAH). First, to learn the compact and useful hash codes from the images, we define a novel loss function for the generator. The loss function mainly consists of classification, similarity, and bits entropy terms. The classification term makes the hash code is discriminative, the similarity term constrains the binary code is similarity preserving, and the bits entropy term assures the learned code is low-error in the quantization. Second, we construct the unique true matrix with the uniform distribution as the input of discriminator to limit the leaned hash codes are bit balanced. The final hash code is learned by a minimax optimization. The positive experimental results on a ground-truth remote sensing image archive validate the usefulness of our GAAH model. Compare with the popular deep hashing methods, our GAAH achieves improved performance.

Many robotic tasks require accurate shape models in order to properly grasp or interact with objects. However, it is often the case that sensors produce incomplete 3D models due to several factors such as occlusion or sensor noise. To address this problem, we propose a semi-supervised method that can recover the complete the shape of a broken or incomplete 3D object model. We formulated a hybrid of 3D variational autoencoder (VAE) and generative adversarial network (GAN) to recover the complete voxelized 3D object. Furthermore, we incorporated a separate classifier in the GAN framework, making it a three player game instead of two which helps stabilize the training of the GAN as well as guides the shape completion process to follow the object class labels. Our experiments show that our model produces 3D object reconstructions with high-similarity to the ground truth and outperforms several baselines in both quantitative and qualitative evaluations.