Analyzing to discover origins of CNNs and ViT architectures in medical images.

In this paper, we introduce in-depth the analysis of CNNs and ViT architectures in medical images, with the goal of providing insights into subsequent research direction. In particular, the origins of deep neural networks should be explainable for medical images, but there has been a paucity of studies on such explainability in the aspect of deep neural network architectures. Therefore, we investigate the origin of model performance, which is the clue to explaining deep neural networks, focusing on the two most relevant architectures, such as CNNs and ViT. We give four analyses, including (1) robustness in a noisy environment, (2) consistency in translation invariance property, (3) visual recognition with obstructed images, and (4) acquired features from shape or texture so that we compare origins of CNNs and ViT that cause the differences of visual recognition performance. Furthermore, the discrepancies between medical and generic images are explored regarding such analyses. We discover that medical images, unlike generic ones, exhibit class-sensitive. Finally, we propose a straightforward ensemble method based on our analyses, demonstrating that our findings can help build follow-up studies. Our analysis code will be publicly available.

Lateral facial thread lifting procedure with temporal anchoring in deep temporal fascia: Anatomical perspectives of superficial temporal artery.

INTRODUCTION: Thread lifting is a non-surgical cosmetic procedure that utilizes threads to lift and tighten sagging skin on the face. In Lateral face lifting with anchoring technique, the threads are inserted into the skin and anchored in place to provide support to the skin at artery free zone. This technique utilizes a long cog thread, allowing for stronger fixation points. The optimal location for thread anchoring is in the fascia of the treatment area. METHOD AND MATERIALS: The study was performed with twelve cadavers with twenty-four specimens of head from cadavers and was processed using phosphotungstic acid-based contrast enhancement micro-computed tomography and conventional computed tomography. The superficial temporal artery with branches of parietal and temporal were then observed with image Slicer program to analyze the safe anchoring place for the deep temporal fascia. The main branch was selected with diameter over 0.3 mm and less than 0.3 mm was regarded as arteriole. Additionally, a case of deep temporal tagging with the Secret Miracle (Hyundae Meditech Co., Ltd., South Korea) has been used for lifting procedures. RESULT: The main branch of the parietal branch located posteriorly was located mean of -13 mm (range of +5.5 mm to -23 mm). And the temporal artery ran most anteriorly had mean of 44 mm anteriorly (range of 32 to 59 mm). The safe area for the tagging is at the deep temporal fascia between the superior temporal line and inferior temporal line. The safe range of deep temporal fascia is a vertical line crossing tragus from 1 to 3 cm anteriorly. CONCLUSION: By analyzing the result of the superficial temporal artery of parietal and temporal branches the ideal tagging place for the thread anchoring area has been suggested.

Clinical Characteristics of Lichen Amyloidosis Associated with Atopic Dermatitis: A Single Center, Retrospective Study.

BACKGROUND: Lichen amyloidosis is a chronic pruritic skin disorder associated with atopic dermatitis, however, the pathogenetic link between these two conditions remains to be elucidated. Only limited research has been performed on patients diagnosed with both pruritic dermatological conditions. OBJECTIVE: This study aimed to analyze the clinical features of lichen amyloidosis associated with atopic dermatitis. METHODS: We conducted a matched case-control study of incident lichen amyloidosis with atopic dermatitis between March 2020 and February 2022. Among the 2,481 patients with atopic dermatitis, 20 patients diagnosed with lichen amyloidosis and atopic dermatitis were included as case patients, and 20 patients diagnosed with atopic dermatitis were enrolled as controls. The controls were matched to cases (1:1) by age and sex. We retrospectively reviewed the medical records of the patients. RESULTS: The prevalence of lichen amyloidosis associated with atopic dermatitis was approximately 0.8%, with a male:female sex ratio of 2.33:1. The recorded onset of lichen amyloidosis associated with atopic dermatitis was more common in adult patients, with moderate-to-severe atopic dermatitis. Lichen amyloidosis lesions in patients with atopic dermatitis were most commonly found on the extremities, sparing the head and neck region. The presence of lichen amyloidosis had no significant impact on severity of atopic dermatitis. CONCLUSION: In patients with lichen amyloidosis associated with atopic dermatitis, the clinical manifestations of lesions are similar to those of conventional lichen amyloidosis lesions in terms of morphology and regional distribution. Further research is required to elucidate the link between the pathogenesis of these two pruritic dermatological conditions.

Deep learning using computed tomography to identify high-risk patients for acute small bowel obstruction: development and validation of a prediction model : a retrospective cohort study.

OBJECTIVE: To build a novel classifier using an optimized 3D-convolutional neural network for predicting high-grade small bowel obstruction (HGSBO). SUMMARY BACKGROUND DATA: Acute SBO is one of the most common acute abdominal diseases requiring urgent surgery. While artificial intelligence and abdominal computed tomography (CT) have been used to determine surgical treatment, differentiating normal cases, HGSBO requiring emergency surgery, and low-grade SBO (LGSBO) or paralytic ileus is difficult. METHODS: A deep learning classifier was used to predict high-risk acute SBO patients using CT images at a tertiary hospital. Images from three groups of subjects (normal, nonsurgical, and surgical) were extracted; the dataset used in the study included 578 cases from 250 normal subjects, with 209 HGSBO and 119 LGSBO patients; over 38 000 CT images were used. Data were analyzed from 1 June 2022 to 5 February 2023. The classification performance was assessed based on accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve. RESULTS: After fivefold cross-validation, the WideResNet classifier using dual-branch architecture with depth retention pooling achieved an accuracy of 72.6%, an area under receiver operating characteristic of 0.90, a sensitivity of 72.6%, a specificity of 86.3%, a positive predictive value of 74.1%, and a negative predictive value of 86.6% on all the test sets. CONCLUSIONS: These results show the satisfactory performance of the deep learning classifier in predicting HGSBO compared to the previous machine learning model. The novel 3D classifier with dual-branch architecture and depth retention pooling based on artificial intelligence algorithms could be a reliable screening and decision-support tool for high-risk patients with SBO.

A Pilot Genome-Wide Association Study Identifies Novel Markers of Metabolic Syndrome in Patients with Psoriasis.

BACKGROUND: Recent studies have reported that psoriasis is associated with the development of metabolic syndrome. Genome-wide association studies have been used to discover gene variant markers that occur frequently in case group in relation to specific diseases. OBJECTIVE: The aim of the present study was to investigate the variants of specific genes involved in metabolic syndrome associated with psoriasis. METHODS: A total of 95 psoriasis patients were recruited and divided into two groups: one with metabolic syndrome (38 patients) and the other without (57 patients). After genotyping, imputation, and quality checking, the association between the several single nucleotide polymorphisms and metabolic syndrome in psoriasis was tested, followed by gene set enrichment analysis. RESULTS: We found 76 gene polymorphisms that conferred an increased risk for metabolic syndrome in patients with psoriasis. Four single nucleotide polymorphisms (rs17154774 of FRMD4A, rs77498336 of GPR116, rs75949580 and rs187682251 of MAPK4) showed the strongest association between metabolic syndrome and psoriasis. The epidermal growth factor receptor protein was located at the center of the protein interactions for the gene polymorphisms. CONCLUSION: This study identified several previously unknown polymorphisms associated with metabolic syndrome in psoriasis. These results highlight the potential for future genetic studies to elucidate the development, and ultimately prevent the onset, of metabolic syndrome in patients with psoriasis.

Multiple Precaching Vehicle Selection Scheme Based on Set Ranking in Intermittently Connected Vehicular Networks.

In vehicular networks, vehicles download vehicular information for various applications, including safety, convenience, entertainment, and social interaction, from the corresponding content servers via stationary roadside units. Since sufficient RSUs might be difficult to deploy due to rough geographical conditions or high deployment costs, vehicular networks can feature uncovered outage zones between two neighboring RSUs. In these outage zones, vehicles cannot download content, and thus the vehicle networks are defined as intermittently connected vehicular networks. In intermittently connected vehicular networks, the download delay and traffic overhead on the backhaul links are increased due to the large size of the content requested by vehicle users and the long distances between RSUs. Using the mobility information of vehicles, several schemes have been proposed to solve this issue by precaching and relaying content via multiple relaying vehicles in the outage zone. However, because they involved the individual ranking of vehicles for precaching and allocated all of the available precaching amounts to the top-ranking vehicles, they decreased the success rate of content requests and the fairness of vehicle precaching. To overcome the problem of these previous schemes, this paper proposes a multiple precaching vehicle selection (MPVS) scheme that efficiently selects a content-precaching vehicle group with multiple precaching vehicles to precache relayed content in outage zones. To achieve this, we first designed numerical models to decide the necessity and the amount of precaching and to calculate the available precaching amounts of vehicles. Next, MPVS calculates all available vehicle sets and ranks each set based on the available precaching amount. Then, the content-precaching vehicle group is identified from the sets by considering both set rankings and vehicle communication overheads. MPVS also provides a content downloading process through the content-precaching vehicle group in the outage zone. Simulation results conducted in various environments with a content request model and a highway mobility model verified that MPVS was superior to a representative previous scheme.

Enhanced Cycling Performance of a Li-Excess Li2CuO2 Cathode Additive by Cosubstitution Nanoarchitectonics of Ni and Mn for Lithium-Ion Batteries.

The adoption of Li2CuO2 has drawn interest as a Li-excess cathode additive for compensating irreversible Li+ loss in anodes during cycling, which would move forward high-energy-density lithium-ion batteries (LIBs). Li2CuO2 provides a high irreversible capacity (>200 mAh g-1) in the first cycle and an operating voltage comparable with commercial cathode materials, but its practical use is still restricted by the structural instability and spontaneous oxygen (O2) evolution, resulting in poor overall cycling performance. It is thus crucial to reinforce the structure of Li2CuO2 to make it more reliable as a cathode additive for charge compensation. Pursuing the structural stability of Li2CuO2, herein, we demonstrate cosubstitution by heteroatoms, such as nickel (Ni) and manganese (Mn), for improving the structural stability and electrochemical performance of Li2CuO2. Such an approach effectively enhances the reversibility of Li2CuO2 by suppressing continuous structural degradation and O2 gas evolution during cycling. Our findings provide new conceptual pathways for developing advanced cathode additives for high-energy LIBs.

Proposal of facial pigmentary unit and facial hyperpigmentation type for Fitzpatrick skin types II-IV.

BACKGROUND: Studies on facial hyperpigmentation across different facial units are limiting. We aimed to analyze melanin pigmentation images to observe facial pigmentary demarcation lines (FPDLs) and suggest facial hyperpigmentation types for normal individuals. MATERIALS AND METHODS: 3D facial melanin pigmentation images of 173 volunteers were obtained and analyzed for the presence of FPDLs. Pigmentation severity was assessed for each of the thirteen facial pigment units. The images were then grouped according to a pattern of hyperpigmentation to suggest three facial hyperpigmentation types-dark spot, photoaging and post-inflammatory hyperpigmentation. RESULTS: Four groups of FPDLs including a novel group I were observed. Nasal, frontal, auricular were the darkest pigmented facial pigment unit, and the anterior neck was the least pigmented. The dark spot type was the most common facial hyperpigmentation type. The photoaging type and the PIH type showed age-dependent distribution, as the photoaging type was more common among the subjects over 40s, and the PIH type was more common in younger subjects. CONCLUSION: Facial hyperpigmentation among healthy individuals with Fitzpatrick skin types II-IV is often accompanied by FPDLs and categorized into three types. Each type is modeled after the pattern of pigmentation associated with certain dermatological disorders. The practical implications of facial hyperpigmentation types can be resourceful in various fields including prevention and treatment of pigmentary disorders.

An analysis of shapes and location of anterior hairline in Asian men.

An individual's first impression can be influenced by a number of esthetic factors, one of which is a natural hairline. The anterior hairline can be affected by various factors, such as race, age, forehead shape, and hair loss progression, similar to androgenetic alopecia. Information on the shape, hairline location on the forehead, and race are required to classify the anterior hairline. According to previous studies, the shape of the anterior hairline can be classified as round, M-shaped, rectangular, bell-shaped, or triangular. This study aimed to analyze the type of anterior hairline in Asian males without androgenetic alopecia. The subjects consisted of 461 males in China, Japan, Kazakhstan, Mongolia, the Philippines, South Korea, and Taiwan. We classified the shape of the anterior hairline into four types: M-shaped (M), rectangular (R), round (O), and wave-shaped (W). We classified the location of the anterior hairline into three types: high (H), middle (M) and low (L), according to the subjects' forehead ratio. Finally, we combined the shape and location types to devise a novel classification system for the anterior hairline type. The most common hairline types in our newly developed classification were MM (23.94%), RM (18.00%) and ML (11.04%). The least common types were OL (0.58%), OH (2.34%) and WH (2.74%). Our results will set standards for hairline shapes and locations, which will be helpful for evaluating hair loss treatment, establishing criteria and designs for reconstructive surgery, and developing camouflage makeups for the forehead hairline.

Optimized Distributed Proactive Caching Based on Movement Probability of Vehicles in Content-Centric Vehicular Networks.

Content-centric vehicular networks (CCVNs) have considered distributed proactive caching as an attractive approach for the timely provision of emerging services. The naïve caching schemes cache all of the contents to only one selected roadside unit (RSU) for requested vehicles to decrease the data acquisition delay between the data source and the vehicles. Due to the high deployment cost for RSUs and their limited capacity of caching, the vehicular networks could support only a limited number of vehicles and a limited amount of content and thus decrease the cache hit ratio. This paper proposes a mobility-aware distributed proactive caching protocol (MDPC) in CCVNs. MDPC caches contents to the selected RSUs according to the movement of vehicles. To reduce the redundancy and the burden of caching for each RSU, MDPC distributes to cache partial contents by the movement pattern, the probability to predict the next locations (RSUs) on the Markov model based on the current RSU. For recovery of prediction failures, MDPC allows each RSU to request partial missing contents to relatively closer neighbor RSUs with a short delay. Next, we expand the protocol with traffic optimization called MDPC_TO to minimize the amount of traffic for achieving proactive caching in CCVNs. In proportion to the mobility probability of a vehicle toward each of the next RSUs, MDPC_TO controls the amount of pre-cached contents in each of the next RSUs. Then, MDPC_TO has constraints to provide enough content from other next RSUs through backhaul links to remove the delay due to prediction failures. Simulation results verify that MDPC_TO produces less traffic than MDPC.

K-Means Clustering-Based Safety System in Large-Scale Industrial Site Using Industrial Wireless Sensor Networks.

A large number of workers and heavy equipment are used in most industrial sizes, and the prevention of safety accidents is one of the most important issues. Therefore, although a number of systems have been proposed to prevent accidents, existing studies assume that workers are gathered in some areas. These assumptions are not suitable for large-scale industrial sites in which workers form as a group and work in a large area. In other words, in a large-scale industrial site, existing schemes are unsuitable for the timely notifying of warnings of threats, and excessive energy is consumed. Therefore, we propose a k-means clustering-based safety system for a large-scale industrial site. In the proposed scheme, workers deployed over a large area are divided into an appropriate number of groups, and threat notification is delivered by a multicasting tree toward each cluster. The notification to workers is delivered through local flooding in each cluster. The simulation results show that the system is able to deliver the notification within a valid time, and it is energy efficient compared to the existing scheme.

Metabolic Responses of Dietary Fiber during Heat Stress: Effects on Reproductive Performance and Stress Level of Gestating Sows.

Heat stress is an important issue, and the addition of fiber to the diet is an option in modifying intestinal health. This study evaluated the effect of acid detergent fiber (ADF) levels on reproductive performance, intestinal integrity, and metabolism of gestating sows, and its carry-over effect on the lactation period during heat stress. The diets included 4.3% (Low fiber; LF), 5.4% (Medium fiber; MF), and 6.5% (High fiber; HF) ADF. Sows fed the HF diet showed a lower respiratory rate, hair cortisol concentration, and farrowing duration compared with the LF treatment. The HF diet increased the pyruvate, citrate cycle, glyoxylate, dicarboxylate, and thiamine metabolism compared with the MF. The concentration of acetate and total short-chain fatty acids were increased in the sows fed the HF diet. The gene expression of glucose transporter 3 and glucose transporter 4 was increased in the HF treatment. The gene expression of heat shock protein 70 was decreased in the HF treatment. The HF diet during gestation increased feed intake, constipation index, piglet weight, and litter weight compared with the LF. Sows in the LF treatment showed the greatest digestibility of crude protein and the lowest digestibility of ADF. In conclusion, a 6.5% ADF level is recommended for gestating sows during heat stress.

Homogenizing Silicon Domains in SiOx Anode during Cycling and Enhancing Battery Performance via Magnesium Doping.

SiOx (x ≈ 1) is one of the most promising anode materials for application in secondary lithium-ion batteries because of its high theoretical capacity. Despite this merit, SiOx has a poor initial Coulombic efficiency, which impedes its widespread use. To overcome this limitation, in this work, we successfully demonstrate a novel synthesis of Mg-doped SiOx via a mass-producible physical vapor deposition method. The solid-state reaction between Mg and SiOx produces Si and electrochemically inert magnesium silicate, thus increasing the initial Coulombic efficiency. The Mg doping concentration determines the phase of the magnesium silicate domains, the size of the Si domains, and the heterogeneity of these two domains. Detailed electron microscopy and synchrotron-based analysis revealed that the nanoscale homogeneity of magnesium silicates driven by cycling significantly affected the lifetime. We found that 8 wt % Mg is the most optimized concentration for enhanced cyclability because MgSiO3, which is the dominant magnesium silicate composition, can be homogeneously mixed with silicon clusters, preventing their aggregation during cycling and suppressing void formation.

Study on Human Activity Recognition Using Semi-Supervised Active Transfer Learning.

In recent years, various studies have begun to use deep learning models to conduct research in the field of human activity recognition (HAR). However, there has been a severe lag in the absolute development of such models since training deep learning models require a lot of labeled data. In fields such as HAR, it is difficult to collect data and there are high costs and efforts involved in manual labeling. The existing methods rely heavily on manual data collection and proper labeling of the data, which is done by human administrators. This often results in the data gathering process often being slow and prone to human-biased labeling. To address these problems, we proposed a new solution for the existing data gathering methods by reducing the labeling tasks conducted on new data based by using the data learned through the semi-supervised active transfer learning method. This method achieved 95.9% performance while also reducing labeling compared to the random sampling or active transfer learning methods.

Energy-Efficient and Reliable Face-Routing Scheme in Wireless Networks.

Face-routing is one of the reliable recovery schemes when geographic routing fails to transmit data packets. Although studies on face-routing can overcome the failure of the data transmission, they lead to much energy consumption due to frequent data transmissions between adjacent nodes for carrying out the rule of face-routing. To avoid the frequent data transmissions, several face-routing schemes have been recently proposed to transmit data packets to the farthest-neighbor node. However, they happen with many data retransmissions because the farthest-neighbor node has a relatively low transmission success ratio. To solve this problem, we propose a new face-routing scheme that determines the most appropriate neighbor node to balance the trade-off between energy efficiency and transmission reliability with two viewpoints. The first viewpoint focuses on how to increase the distance progress of the data delivery in one-hop range to enhance energy efficiency. After that, the second viewpoint focuses on how to increase the success ratio of the data delivery to enhance the transmission reliability. As a result of the simulation, it was confirmed that the proposed method achieves better performance in terms of energy efficiency than existing face-routing research, and it is better than recent face-routing research in terms of reliability and retransmission.

Impact of an anti-Salmonella. Typhimurium Bacteriophage on intestinal microbiota and immunity status of laying hens.

Typhoid is a distinct gastrointestinal disease that largely affects the public by consumption of inadequately or partially cooked eggs from contaminated laying hen farms. This has led the research on laying hens to focus on controlling the contamination by an effective anti-Salmonella spp. agent in the intestine. The treatments included, control, without challenge; PC, Salmonella typhimurium challenged (STC); BP5, 5 ppm bacteriophage/kg + STC; BP10, 10 ppm bacteriophage/kg + STC, on Salmonella shedding, body organs inflammatory reactions, and expression of toll-like receptor (TLR), pro-inflammatory cytokines, and heat shock protein (HSP) in the jejunum, liver,and thigh muscle in the STC laying hens. The RT-PCR method was used to enumerate the number of Salmonella typhimurium in the organs. The birds in the STC groups exhibited the increased population of Salmonella spp. in the excreta (p < .01). In the STC groups, the BP5 and BP10 laying hens exhibited a lower (p < .01) population of Salmonella spp. in the excreta at d 7 after STC. Supplementation of bacteriophage significantly decreased (p < .01) the colonization of S. Typhimurium in the spleen, oviduct, caecum and excreta. Among the STC treatments, the BP10 laying hens showed lower (p < .01) mRNA expression of interferon-γ (IFNγ) and TLR-4 in the jejunum compared with the PC treatment. After the STC, dietary supplementation with BP5 or BP10 decreased (p < .01) the mRNA expressions of IFNγ, HSP-27 and tumour necrosis factor-α in the liver compared with the PC treatment. These results suggest that bacteriophage can be used as an effective agent to decrease S. Typhimurium contamination in laying hens and possibly lower S. Typhimurium transfer to foods.

An Energy Efficient Sink Location Service for Continuous Objects in Wireless Sensor Networks.

In wireless sensor networks (WSNs), detection and report of continuous object, such as forest fire and toxic gas leakage, is one of the major applications. In large-scale continuous object tracking in WSNs, there might be many source nodes simultaneously, detecting the continuous object. Each nodes reports its data to both a base station and mobile workers in the industry field. For communication between the source nodes and a mobile worker, sink location service is needed to continuously notify the location of the mobile worker. But, as the application has a large number of sources, it causes a waste of energy consumption. To address this issue, in this paper, we propose a two-phase sink location service scheme. In the first phase, the proposed scheme constructs a virtual grid structure for merging the source nodes. Then, the proposed scheme aggregates the merging points from an originated merging point as the second phase. Simulation results show that the proposed scheme is superior to other schemes in terms of energy consumption.

Cation-Anion and Anion-CO2 Interactions in Triethyl(octyl)phosphonium Ionic Liquids with Aprotic Heterocyclic Anions (AHAs).

Ionic liquids with aprotic heterocyclic anions (AHAs) have been developed for postcombustion CO2 capture applications. The anions of AHA ILs play a significant role in tuning anion-CO2 complexation. In addition, AHAs are able to trigger the abstraction of acidic protons located at the α position of phosphonium cations by forming hydrogen bonds between cations and anions, eventually leading to cation-driven CO2 complexation. Here we investigate the role of the anion in cation-anion hydrogen bonding and ylide formation. Using CO2 uptake measurements, 31P nuclear magnetic resonance (NMR), attenuated total reflection-Fourier transform infrared (ATR-FTIR) deuterium exchange equilibrium and rates, two-dimensional nuclear Overhauser effect spectroscopy (2D NOESY), and density functional theory calculations, we show that the key is the proximity of the negatively charged nitrogen atoms on the anion to the α protons, which is governed not just by anion basicity but by sterics. Thus, we show that triethyl(octyl)phosphonium 3-methyl-5-trifluoromethylpyrazolide is much more effective in hydrogen-bonding with and deprotonating the cation than the equivalent [P2228] ILs with more basic 2-cyanopyrrolide and 3-trifluoromethylpyrazolide anions.