Visualization Experiments Research Articles

A comparative study of variational autoencoders, normalizing flows, and score-based diffusion models for electrical impedance tomography

Abstract Electrical Impedance Tomography (EIT) is a widely employed imaging technique in industrial inspection, geophysical prospecting, and medical imaging. However, the inherent nonlinearity and ill-posedness of EIT image reconstruction present challenges for classical regularization techniques, such as the critical selection of regularization terms and the lack of prior knowledge. Deep generative models (DGMs) have been shown to play a crucial role in learning implicit regularizers and prior knowledge. This study aims to investigate the potential of three DGMs – variational autoencoder networks, normalizing flow, and score-based diffusion model – to learn implicit regularizers in learning-based EIT imaging. We first introduce background information on EIT imaging and its inverse problem formulation. Next, we propose three algorithms for performing EIT inverse problems based on corresponding DGMs. Finally, we present numerical and visual experiments, which reveal that (1) no single method consistently outperforms the others across all settings, and (2) when reconstructing an object with two anomalies using a well-trained model based on a training dataset containing four anomalies, the conditional normalizing flow (CNF) model exhibits the best generalization in low-level noise, while the conditional score-based diffusion model (CSD*) demonstrates the best generalization in high-level noise settings. We hope our preliminary efforts will encourage other researchers to assess their DGMs in EIT and other nonlinear inverse problems.

Discourse of trauma and transformation of memory in the novel of J. S. Foer’s “Extremely Loud and Incredibly Close”

The article is devoted to considering the peculiarities of trauma discourse in the context of dynamics and transformations of individual and collective memory based on the material of the novel by the modern American writer J. S. Foer “Extremely Loud and Incredibly Close”. Sociocultural dynamics of the first quarter of the 21st century causes significant changes in the spheres of collective imagination and memory, which finds its expression in the literary discourse, one of the significant types of which is currently the discourse of trauma. In turn, the category of trauma acquires universal status in modern humanitarianism. This determines the relevance of the research topic. The purpose of the article is to consider the discursive specificity of J. S. Foer’s novel «Extremely Loud and Incredibly Close» in the context of trauma, individual and collective memory. The subject of consideration is the peculiarities of the representation of trauma in the context of the transformation of memory in the novel by J. S. Foer. The novelty of the study consists in the attempt to distinguish the discourse of trauma related to the peculiarities of the narrator’s vision of reality and the socio-cultural determinations of trauma that affect the entire society, that is, to isolate the universal discourse of social trauma in the novel compared to individual trauma. The task of the research: to analyse the specific features of the discourse of trauma in J. S. Foer’s novel; find out the correlation between trauma discourse and transformations of individual and collective memory; to justify the universal nature of the trauma discourse in the novel. Historical-cultural, hermeneutic and discursive methods were used in the research process. The discourse of trauma in the novel acquires universal features and concerns all the narrators of the novel, who form its plot. The visual component that forms the visual-symbolic narrative of the novel actualizes the synesthetic perception of the trauma and its totality, spread not only to the family of the main character, but also to the entire society. At the same time, the graphic and visual experiments and findings of the author allow us to interpret the text as a complex postmodern and metamodern whole. The trauma discourse in the novel captures the transformations of individual and collective memory, which are aimed at restoring the functionality of the existence of people and communities on both the cognitive and emotional-affective levels. The discourse of trauma acquires features of totality, while remaining rational and avoiding signs of self-obsession, which enables catharsis as life after and with trauma. The symbolic nature of the processes taking place in the collective consciousness and imagination represents itself through the transformations of discourse, which record the transformations of memory. Prospects for further research are outlined by the possibility of considering the discourse of trauma in a wider cultural, linguistic and sociological context.

POOL BOILING HEAT TRANSFER CHARACTERISTICS OF POROUS NICKEL MICROSTRUCTURE SURFACES

Pool boiling is an effective heat dissipation approach in electronic cooling, battery thermal management, etc. This study used the electrochemical deposition method to fabricate one smooth nickel specimen (named Ni-smooth) and three specimens with a porous nickel-stacked structure. The three porous specimens were created with deposition current densities of 0.5 A·cm<sup>-2</sup> (named Ni-0.5), 2.0 A·cm<sup>-2</sup> (names Ni-2.0), and 5.0 A·cm<sup>-2</sup> (named Ni-5.0), respectively. The four samples underwent microstructural characterization via scanning electron microscopy. The increasing current density led to the porous nickel surface exhibiting a more distinct pore structure, and the nickel sphere grains became more refined, developing a loose "mound-like" structure. A marked increase in the nickel film thickness was also observed. Through visual experiments, we evaluated their wettability, and through pool-boiling experiments, we tested their boiling heat-transfer properties. Our findings suggest that samples incorporating a porous nickel structure consistently outperform unmodified samples regarding heat-transfer efficiency. Specifically, sample Ni-0.5A demonstrated the most optimal boiling heat-transfer performance, evidenced by a 32.2% reduction in temperature at the onset of boiling, a 19.9% increase in critical heat flux density, and a 78.6% larger maximum heat-transfer coefficient compared to the smooth nickel sample. These marked improvements are intrinsically linked to the specific characteristics of the porous nickel structure. The higher performance of samples Ni-0.5 can be attributed to the presence of additional nucleation sites within the porous structure and the formation of smaller micro-crystalline dendritic constructs due to the specific current density applied during electrodeposition. Understanding this relationship between surface characteristics and electrodeposition is essential in maximizing heat-transfer efficiency.

Contrastive Registration for Unsupervised Medical Image Segmentation.

Medical image segmentation is an important task in medical imaging, as it serves as the first step for clinical diagnosis and treatment planning. While major success has been reported using deep learning supervised techniques, they assume a large and well-representative labeled set. This is a strong assumption in the medical domain where annotations are expensive, time-consuming, and inherent to human bias. To address this problem, unsupervised segmentation techniques have been proposed in the literature. Yet, none of the existing unsupervised segmentation techniques reach accuracies that come even near to the state-of-the-art of supervised segmentation methods. In this work, we present a novel optimization model framed in a new convolutional neural network (CNN)-based contrastive registration architecture for unsupervised medical image segmentation called CLMorph. The core idea of our approach is to exploit image-level registration and feature-level contrastive learning, to perform registration-based segmentation. First, we propose an architecture to capture the image-to-image transformation mapping via registration for unsupervised medical image segmentation. Second, we embed a contrastive learning mechanism in the registration architecture to enhance the discriminative capacity of the network at the feature level. We show that our proposed CLMorph technique mitigates the major drawbacks of existing unsupervised techniques. We demonstrate, through numerical and visual experiments, that our technique substantially outperforms the current state-of-the-art unsupervised segmentation methods on two major medical image datasets.

A study of visual attention patterns of snow and ice athletes based on eye-tracking technology

Abstract The visual system has a strong information processing ability, and visual attention tracking has various applications in various scenes. This paper mainly focuses on the sports scene of ice and snow far mobilization. It constructs a visual attention system model based on eye tracking. It first establishes an eye tracking system framework using deep learning, and improves the gaze estimation by optimizing the feature extraction network. The visual attention system model was constructed using particle filtering based on motion feature cognition. In the eye-tracking visual attention system model experiments, the Accuracy of the improved eye-tracking system in this paper can be significantly improved to 1.13°, and the error of the visual attention system can be kept within 10°. Furthermore, the four ice and snow sports scene types have an average accuracy of 85.47%, and the constructed model performs well. This study offers a guide for effectively combining eye tracking technology and visual attention.

Innovative Research on Illustration Design Integrating Color Science and Image Processing Technology

Abstract How to create intelligently and efficiently has become a hot research topic in the field of illustration design. In this paper, starting from the acceptance intention to satisfy users’ favorites and needs, we propose an intelligent analysis method of style based on a deep clustering model and use the categorized style attributes as a reference to disperse designers’ creative thinking. In addition, a semi-supervised semantic segmentation method is introduced to construct an automatic coloring model for sketches, and the automatic coloring function is realized by incorporating the color label recognition process, which further improves the efficiency of designers’ work. The results of clustering analysis of the STL dataset show that the ACC (78.39%), NMI (67.78%), and ARI (62.93%) metrics of the VIT+K-Means model have the largest results compared with other feature extraction methods. Results are superior to other feature extraction methods. Not only that, adding a pseudo-label retraining process for this model further improves the results of the three metrics by 9.13%, 9.97%, and 10.78%, and the visual analysis experiments of clustering clusters also verified the performance enhancement effect. In the comparative analysis of illustration coloring models, the AdvSSL scheme with semi-supervised strategy achieves the best performance in the semantic segmentation task, with an improvement of 4.31% and 2.29% over the SSAN scheme and the S4GCN scheme, respectively. The experts’ coloring evaluation shows that the improved sketch coloring model has the highest results in all four dimensions compared to the traditional Tag2Pix model.

Impact of energy efficiency information on homebuying searches: evidence from a visual choice experiment in the USA

Impact of energy efficiency information on homebuying searches: evidence from a visual choice experiment in the USA

Visuomotor Navigation for Embodied Robots With Spatial Memory and Semantic Reasoning Cognition.

The fundamental prerequisite for embodied agents to make intelligent decisions lies in autonomous cognition. Typically, agents optimize decision-making by leveraging extensive spatiotemporal information from episodic memory. Concurrently, they utilize long-term experience for task reasoning and foster conscious behavioral tendencies. However, due to the significant disparities in the heterogeneous modalities of these two cognitive abilities, existing literature falls short in designing effective coupling mechanisms, thus failing to endow robots with comprehensive intelligence. This article introduces a navigation framework, the hierarchical topology-semantic cognitive navigation (HTSCN), which seamlessly integrates both memory and reasoning abilities within a singular end-to-end system. Specifically, we represent memory and reasoning abilities with a topological map and a semantic relation graph, respectively, within a unified dual-layer graph structure. Additionally, we incorporate a neural-based cognition extraction process to capture cross-modal relationships between hierarchical graphs. HTSCN forges a link between two different cognitive modalities, thus further enhancing decision-making performance and the overall level of intelligence. Experimental results demonstrate that in comparison to existing cognitive structures, HTSCN significantly enhances the performance and path efficiency of image-goal navigation. Visualization and interpretability experiments further corroborate the promoting role of memory, reasoning, as well as their online learned relationships, on intelligent behavioral patterns. Furthermore, we deploy HTSCN in real-world scenarios to further verify its feasibility and adaptability.

CHARACTERIZATION OF SEPARATED SHEAR LAYER OF ELEVATED JET IN CROSSFLOW AT LOW-VELOCITY RATIO

The present study reports an experimental investigation of the flow dynamics of separated shear layers of an elevated jet-in-crossflow (EJICF). The current work aims to enhance the understanding of the downwash of jet fluid into the stack-wake region, akin to plume downwash in chimney flows, particularly at a low-velocity ratio. One end of the square stack with the aspect ratio of 9.0 is kept on a flat plate while the jet emerges from the free-end open to crossflow. Flow visualization and particle image velocimetry (PIV) experiments are conducted in a recirculating, low-speed water tunnel for three Reynolds numbers, namely Re = 1000, 2000, and 4000, based on the side of the square stack (<i>D</i>) and the freestream velocity (<i>U</i><sub>∞</sub>). The study mainly focuses on the jet shear layer characteristics and the downstream wake near the free-end formed by the interaction of crossflow and the jet shear layer. The instantaneous flow behavior includes the analysis of the shedding of the jet shear layer in a symmetric plane (<i>z</i> = 0) at a constant velocity ratio of <i>R</i> = 0.5. On the windward side, the jet shear layer reveals the formation of small-scale vortices due to the Kelvin-Helmholtz (KH) instability for Re ≥ 2000. Also, it is found that at a high Reynolds number (Re = 4000), an unstable vortex appears above the jet exit on the upstream side, mentioned as a hovering vortex. The mean and fluctuating flow behavior are investigated using streamline and contour plots of mean velocity and fluctuating components.

Lighting preferences of Iberian sausages for Spanish observers

We report results of a visual experiment performed by 20 adult Spanish observers using calibrated images of 18 samples of sliced Iberian sausages under 15 illuminations, including 8 illuminants proposed by the International Commission on Illumination (CIE) and 7 light sources found in meat sections of supermarkets in Spain and Japan. Visual results from the pair comparison and magnitude estimation psychophysical methods used in our experiment showed a high correlation (Pearson’s linear correlation coefficient r = 0.915). Illumination strongly influenced color preferences of sliced Iberian sausages. For example, using the magnitude estimation method, the average lighting preferences changed about 2.8/10 for the 15 illuminations tested. There were no statistically significant differences between lighting preferences for different kind of sausages. Exploring quality indices of white lights, we found that the correlated color temperature and gamut area index (Color Res. Appl. 33, 192-202, 2008) provided the highest linear correlations with visual preferences in our current experiment. Lights with correlated color temperatures below 3000 K or gamut area indices below 70 were not preferred. The CIE general color rendering and color fidelity indices provided poor predictions of visual preferences in our current experiment.

Delineating the effects of morphological changes on retinal hemodynamics in diabetic human retinas: An invitro investigation

Diabetes-related vascular abnormalities are commonly encountered in the retinal vasculature, the most common being the increased tortuosity and reduced lumen diameter. These morphological changes significantly affect the underlying hemodynamics and remain unexplored. The current study examines the changes produced by the altered morphology on the blood flow velocity and wall shear stress (WSS) using particle streak velocimetry (PSV) experiments and numerical simulations and compares it with the healthy scenario. The diabetic and healthy retinal arteries are extracted from patient-specific fundus images and transparently printed using stereolithography (SLA) in three dimensions. Flow visualization experiments are performed using the blood-mimicking fluid (BMF) with three different flow rates, whose values are evaluated using the similarity analysis. Experimental and simulation results illustrate a significant reduction in the average axial velocity and WSS due to tortuosity with maximum percentage reductions of 10% and 11%, respectively, compared to healthy. Additionally, a strong correlation is observed between a particular artery's local tortuosity and the WSS, with a higher tortuosity index leading to the development of low-shear regions inside the diabetic artery. A 14% diameter reduction in the diabetic arteries leads to approximately 65% and 63% increase in the velocity and WSS, respectively, compared to healthy. Moreover, the experiments also capture the effect of decreased blood viscosity on the WSS in the case of diabetic arteries, illustrating a significant decrease in WSS values. The experiments performed in the current study may help ophthalmologists estimate the induced damage that the arterial morphological changes may create and facilitate an accurate prognosis of retinal vascular diseases.

BUBBLE DEFECTS AND CONTROL OF THERMAL CURING OF WATERBORNE RESIN COATINGS

To solve the problem of bubble retention leading to a decrease in the performance of waterborne resin coatings during thermal curing, a rapid thermal curing coating based on an aqueous epoxy-modified acrylic resin containing an amino resin cross linker was synthesized through solution copolymerization. Using a synchronous thermal analyzer, the mass loss and heat flow during the thermal curing process of the coating was measured. In addition, an in situ thermal curing visualization system was used to observe the curing behavior at the easy-open end notches under different heating rates. The results indicated that the curing process is divided into an initial rapid heating stage and a subsequent slower heating stage by the boiling point of water, with a critical temperature range of 126-150°C. The visualization experiments showed that when the temperature reached 100°C, the water quickly evaporated to form bubbles. Increasing the heating rate before this temperature caused the bubbles to rapidly escape. Therefore, an optimal heating profile with a rapid initial heating rate of 3.13°C/s, followed by a slower heating rate of 0.52°C/s in the next stage, is proposed. This discovery is of great significance for optimizing the thermal curing process of waterborne coatings on metal substrates, including those used for easy-open ends.

Semantic and Correlation Disentangled Graph Convolutions for Multilabel Image Recognition.

Multilabel image recognition (MLR) aims to annotate an image with comprehensive labels and suffers from object occlusion or small object sizes within images. Although the existing works attempt to capture and exploit label correlations to tackle these issues, they predominantly rely on global statistical label correlations as prior knowledge for guiding label prediction, neglecting the unique label correlations present within each image. To overcome this limitation, we propose a semantic and correlation disentangled graph convolution (SCD-GC) method, which builds the image-specific graph and employs graph propagation to reason the labels effectively. Specifically, we introduce a semantic disentangling module to extract categorywise semantic features as graph nodes and develop a correlation disentangling module to extract image-specific label correlations as graph edges. Performing graph convolutions on this image-specific graph allows for better mining of difficult labels with weak visual representations. Visualization experiments reveal that our approach successfully disentangles the dominant label correlations existing within the input image. Through extensive experimentation, we demonstrate that our method achieves superior results on the challenging Microsoft COCO (MS-COCO), PASCAL visual object classes (PASCAL-VOC), NUS web image dataset (NUS-WIDE), and Visual Genome 500 (VG-500) datasets. Code is available at GitHub: https://github.com/caigitrepo/SCDGC.

Experimental study and modeling of water film thickness in aero-engine under air–water mist flow

The water film not only plays an important role in the mass, momentum, and energy transfer between the air–water-surface but also determines the on-wing washing effect of the aero-engine. In view of this, air–water mist flow visualization experiments have been conducted at different gas velocities and water-to-air ratios in a compressor cascade, and the microscopic water film images have been analyzed to extract the transient water film thickness data by the Matlab code. It was found that the transient water film thickness fluctuation has no obvious association with the gas velocity, and the water film fluctuation is more affected by the water-to-air ratio. As the water-to-air ratio increases, the fluctuation magnitude of the water film thickness increases. The average water film thickness has been studied in relation to gas velocity and water flow rate, i.e., the average thickness of water film decreases with increasing gas velocity and increase with the increasing water flow rate. On the basis of the water film flow equation and taking the droplet collection efficiency into account, i.e., from the perspective of the physical mechanism of water film formed, a new model for predicting the water film thickness of a compressor blade surface under the air–water mist flow condition was proposed and validated. This model predicted, with a root mean square error and the mean absolute percentage error of 11.6% and 9.15%, respectively, under the present experimental flow conditions.

Thermal phenomena and size effects of Mg powder in combustion process.

Magnesium is a combustible metal that poses various safety risks, including fires and explosions. However, there are limited safety measures available to prevent and respond to potential fires and explosion incidents in the metal industry. In this study, the combustion process of Mg fires was closely examined using infrared thermal imaging, focusing on the effects of Mg powder size. For the experiment, Mg powder was burned by increasing the temperature to approximately 967.4 K using an ignition unit and controller equipped with a tungsten heater. Moreover, combustion velocity measurement experiments for Mg particle sizes of 75, 100, and 150 μm were conducted using the combustion velocity measurement device presented in the NFPA 484 standard. On combustion of Mg, flames are observed; smoke is emitted as demonstrated by thermal and flow visualization experiments. The combustion velocity measurement experiment results demonstrated that the greater the slope value (combustion velocity) for the combustion length over time, the faster is the combustion velocity, with the 75 μm particle size having the fastest combustion velocity. The results of this experiment can be utilized as references for Mg fire control design and to gain a better understanding of the scope of smoke and fire hazard investigation measures.

Optimizing user experience in SSVEP-BCI systems.

The emergence of brain-computer interface (BCI) technology provides enormous potential for human medical and daily applications. Therefore, allowing users to tolerate the visual response of SSVEP for a long time has always been an important issue in the SSVEP-BCI system. We recruited three subjects and conducted visual experiments in groups using different frequencies (17 and 25Hz) and 60Hz light. After recording the physiological signal, use FFT to perform a time-frequency analysis on the physiological signal to check whether there is any difference in the signal-to-noise ratio and amplitude of the 60Hz light source compared with a single low-frequency signal source. The results show that combining a 60Hz light source with low-frequency LEDs can reduce participants' eye discomfort while achieving effective light stimulation control. At the same time, there was no significant difference in signal-to-noise ratio and amplitude between the groups. This also means that 60Hz can make vision more continuous and improve the subject's experience and comfort. At the same time, it does not affect the performance of the original SSVEP-induced response. This study highlights the importance of considering technical aspects and user comfort when designing SSVEP-BCI systems to increase the usability of SSVEP systems for long-term flash viewing.

VISUALIZATION OF MIXING UPPER AND LOWER STRATIFIED SUSPENSIONS USING RED AND BLUE FLUORESCENT PARTICLES

We conducted visualization experiments to investigate the macroscopic mixing of upper and lower stratified suspensions. We used spectroscopy to distinguish the upper and lower suspensions in the images recorded by a color camera, in which the upper and lower suspended particles were replaced with blue and red fluorescent particles, respectively. We also proposed a method to determine the mass of adsorbed particles and the volume of fluorescent solution required to produce fluorescent particles by using the two calibration curves and the Freundlich equation. Visualization with fluorescent particles proved to be effective for macroscopic mixing by comparing mono and color images. Finally, we found good agreement in the trends between the experimental and numerical results.

GLO GERM AND COVID-19: ILLUMINATING HYGIENE AND PROMOTING TRANSMISSION AWARENESS.

Maintaining effective surface hygiene and preventing contamination is of paramount importance. Our study introduces Glo Germ, a versatile product available in various forms, which possesses the unique ability to reveal hidden truths under ultraviolet light, enhance understanding of hygiene, and spread awareness of COVID-19 transmission and preventive measures. A comprehensive study was conducted to assess different surface cleaning techniques' effectiveness. Glo Germ, containing a fluorescent dye activated by ultraviolet light, was used to visualize germ spread and compare disinfectant cleaners' efficacy. The study encompassed diverse surfaces and materials, aiming to identify optimal cleaning techniques for each context. Furthermore, a small illustrative study was conducted during a COVID-19 awareness presentation involving students. Glo Germ was applied to hands, revealing its subsequent spread to faces and surfaces. This visual experiment effectively emphasized hand hygiene and mask-wearing importance. Results indicated that while water alone achieved satisfactory cleaning results, using detergent and the appropriate cleaning tools further improved efficacy. Notably, adhering to consistent patterns and applying pressure during cleaning proved essential. The student demonstration showed how contaminants spread quickly, highlighting hand hygiene's significance and the potential extent of contamination through sneezing. Glo Germ inclusion in these experiments highlights its potential in educating about surface cleaning and microbial transmission, offering an interactive and engaging approach to promoting personal hygiene and fostering illness prevention awareness.

The Rational Agent Benchmark for Data Visualization.

Understanding how helpful a visualization is from experimental results is difficult because the observed performance is confounded with aspects of the study design, such as how useful the information that is visualized is for the task. We develop a rational agent framework for designing and interpreting visualization experiments. Our framework conceives two experiments with the same setup: one with behavioral agents (human subjects), and the other one with a hypothetical rational agent. A visualization is evaluated by comparing the expected performance of behavioral agents to that of a rational agent under different assumptions. Using recent visualization decision studies from the literature, we demonstrate how the framework can be used to pre-experimentally evaluate the experiment design by bounding the expected improvement in performance from having access to visualizations, and post-experimentally to deconfound errors of information extraction from errors of optimization, among other analyses.

Experimental and modeling studies on continuous liquid removal in horizontal gas wells

The potential risks of liquid-loading can be significantly decreased by precisely calculating the minimum gas flowrate required for continuous liquid removal in gas wells and using suitable deliquification technology beforehand. Due to lack of comparative studies with liquid-loading characteristics, existing prediction models are not very adaptable in the course of the application. So as to investigate the flowing behavior of liquid film under different conditions, visual experiment was conducted. The findings indicate that as the inclined angle increases, the liquid-film-reversal gas velocity increases initially before decreasing. The maximum velocity of liquid-film-reversal is around 55°. The liquid-film-reversing gas velocity increases linearly along with the rise in superficial liquid velocities. The liquid-film-reversal gas velocity likewise increases linearly as the superficial liquid velocities. The analysis findings also indicate that several dynamic liquid-loading symptoms of the gas well are inconsistent with the liquid-film-reversal criterion, meaning that the gas well does not instantly follow the liquid-film-reversal. On the basis of our experimental findings, a new liquid-loading commencement criteria was then developed. As a result, this research suggests a novel model for evaluating liquid-loading in gas wells. The model’s accuracy was found to be as high as 85.7% by looking at 14 gas wells and perform better than other models in the Coleman dataset, which can theoretically enable the prediction of liquid-loading in gas wells.