RESUMO
The fovea is a small region within the central retina that is responsible for our high acuity daylight vision. Chickens also have a high acuity area (HAA), and are one of the few species that enables studies of the mechanisms of HAA development, due to accessible embryonic tissue and methods to readily perturb gene expression. To enable such studies, we characterized the development of the chick HAA using single molecule fluorescent in situ hybridization (smFISH), along with more classical methods. We found that Fgf8 provides a molecular marker for the HAA throughout development and into adult stages, allowing studies of the cellular composition of this area over time. The radial dimension of the ganglion cell layer (GCL) was seen to be the greatest at the HAA throughout development, beginning during the period of neurogenesis, suggesting that genesis, rather than cell death, creates a higher level of retinal ganglion cells (RGCs) in this area. In contrast, the HAA acquired its characteristic high density of cone photoreceptors post-hatching, which is well after the period of neurogenesis. We also confirmed that rod photoreceptors are not present in the HAA. Analyses of cell death in the developing photoreceptor layer, where rods would reside, did not show apoptotic cells, suggesting that lack of genesis, rather than death, created the "rod-free zone" (RFZ). Quantification of each cone photoreceptor subtype showed an ordered mosaic of most cone subtypes. The changes in cellular densities and cell subtypes between the developing and mature HAA provide some answers to the overarching strategy used by the retina to create this area and provide a framework for future studies of the mechanisms underlying its formation.
Assuntos
Retina , Células Ganglionares da Retina , Animais , Embrião de Galinha , Células Ganglionares da Retina/citologia , Retina/embriologia , Células Fotorreceptoras Retinianas Cones/metabolismo , Galinhas , Neurogênese/fisiologia , Fator 8 de Crescimento de Fibroblasto/metabolismo , Fator 8 de Crescimento de Fibroblasto/genética , Hibridização in Situ Fluorescente , Fóvea Central/embriologia , Acuidade Visual , Células Fotorreceptoras Retinianas Bastonetes/metabolismo , Células Fotorreceptoras Retinianas Bastonetes/citologia , Regulação da Expressão Gênica no DesenvolvimentoRESUMO
The vertebrate retina is generated by retinal progenitor cells (RPCs), which produce >100 cell types. Although some RPCs produce many cell types, other RPCs produce restricted types of daughter cells, such as a cone photoreceptor and a horizontal cell (HC). We used genome-wide assays of chromatin structure to compare the profiles of a restricted cone/HC RPC and those of other RPCs in chicks. These data nominated regions of regulatory activity, which were tested in tissue, leading to the identification of many cis-regulatory modules (CRMs) active in cone/HC RPCs and developing cones. Two transcription factors, Otx2 and Oc1, were found to bind to many of these CRMs, including those near genes important for cone development and function, and their binding sites were required for activity. We also found that Otx2 has a predicted autoregulatory CRM. These results suggest that Otx2, Oc1 and possibly other Onecut proteins have a broad role in coordinating cone development and function. The many newly discovered CRMs for cones are potentially useful reagents for gene therapy of cone diseases.
Assuntos
Dissecação , Fator 6 Nuclear de Hepatócito/metabolismo , Fatores de Transcrição Otx/metabolismo , Retina/crescimento & desenvolvimento , Células Fotorreceptoras Retinianas Cones/metabolismo , Fatores de Transcrição/metabolismo , Animais , Sítios de Ligação , Galinhas , Cromatina , Regulação da Expressão Gênica no Desenvolvimento , Redes Reguladoras de Genes , Fator 6 Nuclear de Hepatócito/genética , Fatores de Transcrição Otx/genética , Retina/metabolismo , Células-TroncoRESUMO
Concomitant activation of the Wnt pathway and suppression of Mapk signalling by two small molecule inhibitors (2i) in the presence of leukaemia inhibitory factor (LIF) (hereafter termed 2i/L) induces a naive state in mouse embryonic stem (ES) cells that resembles the inner cell mass (ICM) of the pre-implantation embryo. Since the ICM exists only transiently in vivo, it remains unclear how sustained propagation of naive ES cells in vitro affects their stability and functionality. Here we show that prolonged culture of male mouse ES cells in 2i/L results in irreversible epigenetic and genomic changes that impair their developmental potential. Furthermore, we find that female ES cells cultured in conventional serum plus LIF medium phenocopy male ES cells cultured in 2i/L. Mechanistically, we demonstrate that the inhibition of Mek1/2 is predominantly responsible for these effects, in part through the downregulation of DNA methyltransferases and their cofactors. Finally, we show that replacement of the Mek1/2 inhibitor with a Src inhibitor preserves the epigenetic and genomic integrity as well as the developmental potential of ES cells. Taken together, our data suggest that, although short-term suppression of Mek1/2 in ES cells helps to maintain an ICM-like epigenetic state, prolonged suppression results in irreversible changes that compromise their developmental potential.
Assuntos
Células-Tronco Embrionárias/citologia , Células-Tronco Embrionárias/enzimologia , MAP Quinase Quinase 1/antagonistas & inibidores , MAP Quinase Quinase 2/antagonistas & inibidores , Animais , Blastocisto , Instabilidade Cromossômica , Metilação de DNA , Feminino , Impressão Genômica , Cariotipagem , Masculino , CamundongosRESUMO
Monocular depth estimation is a task aimed at predicting pixel-level distances from a single RGB image. This task holds significance in various applications including autonomous driving and robotics. In particular, the recognition of surrounding environments is important to avoid collisions during autonomous parking. Fisheye cameras are adequate to acquire visual information from a wide field of view, reducing blind spots and preventing potential collisions. While there have been increasing demands for fisheye cameras in visual-recognition systems, existing research on depth estimation has primarily focused on pinhole camera images. Moreover, depth estimation from fisheye images poses additional challenges due to strong distortion and the lack of public datasets. In this work, we propose a novel underground parking lot dataset called JBNU-Depth360, which consists of fisheye camera images and their corresponding LiDAR projections. Our proposed dataset was composed of 4221 pairs of fisheye images and their corresponding LiDAR point clouds, which were obtained from six driving sequences. Furthermore, we employed a knowledge-distillation technique to improve the performance of the state-of-the-art depth-estimation models. The teacher-student learning framework allows the neural network to leverage the information in dense depth predictions and sparse LiDAR projections. Experiments were conducted on the KITTI-360 and JBNU-Depth360 datasets for analyzing the performance of existing depth-estimation models on fisheye camera images. By utilizing the self-distillation technique, the AbsRel and SILog error metrics were reduced by 1.81% and 1.55% on the JBNU-Depth360 dataset. The experimental results demonstrated that the self-distillation technique is beneficial to improve the performance of depth-estimation models.
RESUMO
Even though plastic improved the human standard of living, handling the plastic waste represents an enormous challenge. It takes more than 100 years to decompose discarded or buried waste plastics. Microplastics are one of the causes of significantly pervasive environmental pollutants. The incineration of plastic waste generates toxic gases, underscoring the need for new approaches, in contrast to conventional strategies that are required for recycling plastic waste. Therefore, several studies have attempted to upcycle plastic waste into high value-added products. Converting plastic waste into carbonaceous materials is an excellent upcycling technique due to their diverse practical applications. This review summarizes various studies dealing with the upcycling of plastic waste into carbonaceous products. Further, this review discusses the applications of carbonaceous products synthesized from plastic waste including carbon fibers, absorbents for water purification, and electrodes for energy storage. Based on the findings, future directions for effective upcycling of plastic waste into carbonaceous materials are suggested.
Assuntos
Plásticos , Reciclagem , Gases , HumanosRESUMO
The conventional finger-vein recognition system is trained using one type of database and entails the serious problem of performance degradation when tested with different types of databases. This degradation is caused by changes in image characteristics due to variable factors such as position of camera, finger, and lighting. Therefore, each database has varying characteristics despite the same finger-vein modality. However, previous researches on improving the recognition accuracy of unobserved or heterogeneous databases is lacking. To overcome this problem, we propose a method to improve the finger-vein recognition accuracy using domain adaptation between heterogeneous databases using cycle-consistent adversarial networks (CycleGAN), which enhances the recognition accuracy of unobserved data. The experiments were performed with two open databases-Shandong University homologous multi-modal traits finger-vein database (SDUMLA-HMT-DB) and Hong Kong Polytech University finger-image database (HKPolyU-DB). They showed that the equal error rate (EER) of finger-vein recognition was 0.85% in case of training with SDUMLA-HMT-DB and testing with HKPolyU-DB, which had an improvement of 33.1% compared to the second best method. The EER was 3.4% in case of training with HKPolyU-DB and testing with SDUMLA-HMT-DB, which also had an improvement of 4.8% compared to the second best method.
Assuntos
Dedos , Veias , Bases de Dados Factuais , Hong Kong , HumanosRESUMO
Among many available biometrics identification methods, finger-vein recognition has an advantage that is difficult to counterfeit, as finger veins are located under the skin, and high user convenience as a non-invasive image capturing device is used for recognition. However, blurring can occur when acquiring finger-vein images, and such blur can be mainly categorized into three types. First, skin scattering blur due to light scattering in the skin layer; second, optical blur occurs due to lens focus mismatching; and third, motion blur exists due to finger movements. Blurred images generated in these kinds of blur can significantly reduce finger-vein recognition performance. Therefore, restoration of blurred finger-vein images is necessary. Most of the previous studies have addressed the restoration method of skin scattering blurred images and some of the studies have addressed the restoration method of optically blurred images. However, there has been no research on restoration methods of motion blurred finger-vein images that can occur in actual environments. To address this problem, this study proposes a new method for improving the finger-vein recognition performance by restoring motion blurred finger-vein images using a modified deblur generative adversarial network (modified DeblurGAN). Based on an experiment conducted using two open databases, the Shandong University homologous multi-modal traits (SDUMLA-HMT) finger-vein database and Hong Kong Polytechnic University finger-image database version 1, the proposed method demonstrates outstanding performance that is better than those obtained using state-of-the-art methods.
Assuntos
Biometria , Veias , Dedos/diagnóstico por imagem , Hong Kong , Humanos , Movimento (Física)RESUMO
The resolution of Si complementary metal-oxide-semiconductor field-effect transistor (C-MOSFET) image sensors (CISs) has been intensively enhanced to follow the technological revolution of smartphones, AI devices, autonomous cars, robots, and drones, approaching the physical and material limits of a resolution increase in conventional Si CISs because of the low quantum efficiency (i.e., ~40%) and aperture ratio (i.e., ~60%). As a novel solution, a hybrid organic-Si image sensor was developed by implementing B, G, and R organic photodiodes on four n-MOSFETs for photocurrent sensing. Photosensitive organic donor and acceptor materials were designed with cost-effective small molecules, i.e., the B, G, and R donor and acceptor small molecules were Coumarin6 and C_60, DMQA and MePTC, and ZnPc and TiOPc, respectively. The output voltage sensing margins (i.e., photocurrent signal difference) of the hybrid organic-Si B, G, and R image sensor pixels presented results 17, 11, and 37% higher than those of conventional Si CISs. In addition, the hybrid organic-Si B, G, and R image sensor pixels could achieve an ideal aperture ratio (i.e., ~100%) compared with a Si CIS pixel using the backside illumination process (i.e., ~60%). Moreover, they may display a lower fabrication cost than image sensors because of the simple image sensor structure (i.e., hybrid organic-Si photodiode with four n-MOSFETs).
RESUMO
The present protocol describes the creation of four-dimensional (4D), time-dependent, shape-changeable, stimuli-responsive soft robots using a three-dimensional (3D) bio-printing method. Recently, 4D printing techniques have been extensively proposed as innovative new methods for developing shape-transformable soft robots. In particular, 4D time-dependent shape transformation is an essential factor in soft robotics because it allows effective functions to occur at the right time and place when triggered by external cues, such as heat, pH, and light. In line with this perspective, stimuli-responsive materials, including hydrogels, polymers, and hybrids, can be printed to realize smart shape-transformable soft robotic systems. The current protocol can be used to fabricate thermally responsive soft grippers composed of N-isopropylacrylamide (NIPAM)-based hydrogels, with overall sizes ranging from millimeters to centimeters in length. It is expected that this study will provide new directions for realizing intelligent soft robotic systems for various applications in smart manipulators (e.g., grippers, actuators, and pick-and-place machines), healthcare systems (e.g., drug capsules, biopsy tools, and microsurgeries), and electronics (e.g., wearable sensors and fluidics).
Assuntos
Robótica , Polímeros Responsivos a Estímulos , Hidrogéis , Robótica/métodos , Polímeros , Impressão TridimensionalRESUMO
Fumed silica-based ultra-high-purity synthetic quartz powder was developed via the sol-gel process to apply to quartz wares and quartz crucibles for use in advanced semiconductor processes. The process conditions of preparing potassium silicate solution, gelation, and cleaning were optimized, i.e., the relative ratio of fumed silica (10 wt%) to KOH (4 wt%) for potassium silicate solution, gelation time 3 h, and cleaning for 1 h with 5 wt% HCl solution. It was observed that the gelation time strongly affected the size distribution of the quartz powder; i.e., a longer gelation time led to a larger size (d50) of the synthesized quartz powder: 157 µm for 2 h and 331 µm for 5 h. In particular, it was found that the morphology of the as-synthesized quartz powder greatly depended on the pulverizing process; i.e., the shape of quartz powder was shown to be rod-shaped for the without-gel-pulverizing process and granular-shaped with the process. We expect that the fumed silica-based ultra-high-purity quartz powder with an impurity level of 74.1 ppb synthesized via the sol-gel process is applicable as a raw material for quartz wares and crucibles for advanced semiconductor processes beyond the design rule of 3 nm.
RESUMO
Purpose: An increasing number of older patients now undergo liver transplantation (LT). Although the overall outcomes in older patients are not different from those of younger patients, there is no tool to predict LT prognosis in older patients. We hypothesized that a modified Charlson comorbidity index (mCCI) and 5-factor modified frailty index (mFI-5) can predict outcomes in older patients after LT. Methods: This retrospective study included 155 patients (aged >65 years) who underwent LT at Seoul National University Hospital. The recipients were subcategorized into 2 groups based on the mCCI score and mFI-5: the low (0-1) and high (2-5) mCCI groups, and low (≤0.4) and high (>0.4) mFI-5 groups. The independent effect of each variable on post-LT survival was determined using the mCCI subgroup, age at transplantation, sex, Child-Turcotte-Pugh score, model for end-stage liver disease (MELD) score, and mFI-5 subgroup. Results: The high-mCCI group (41 patients) showed significantly lower 1- and 3-month and 1-, 3-, and 5-year survival than the low-mCCI group. Using the Cox regression model, the mCCI, sex, and MELD score remained significant. The mFI-5 was not a significant factor to predict patients' survival. Conclusion: The mCCI and MELD scores could be used to predict post-LT survival in older patients.
RESUMO
The cyanation of aromatic boronic acids, boronate esters, and borate salts was developed under copper-mediated oxidative conditions using ammonium iodide and DMF as the source of nitrogen and carbon atom of the cyano unit, respectively. The procedure was successfully extended to the cyanation of electron-rich benzenes, and regioselective introduction of a cyano group at the arene C-H bonds was also achieved. The observation that the reaction proceeds via a two-step process, initial iodination and then cyanation, led us to propose that ammonium iodide plays a dual role to provide iodide and nitrogen atom of the cyano moiety.
Assuntos
Boratos/química , Ácidos Borônicos/química , Cobre/química , Ésteres/química , Formamidas/química , Iodetos/química , Nitrilas/síntese química , Compostos de Amônio Quaternário/química , Dimetilformamida , Estrutura Molecular , Nitrilas/química , EstereoisomerismoRESUMO
Morphological attributes of human blastocyst components and their characteristics are highly correlated with the success rate of in vitro fertilization (IVF). Blastocyst component analysis aims to choose the most viable embryos to improve the success rate of IVF. The embryologist evaluates blastocyst viability by manual microscopic assessment of its components, such as zona pellucida (ZP), trophectoderm (TE), blastocoel (BL), and inner cell mass (ICM). With the success of deep learning in the medical diagnosis domain, semantic segmentation has the potential to detect crucial components of human blastocysts for computerized analysis. In this study, a sprint semantic segmentation network (SSS-Net) is proposed to accurately detect blastocyst components for embryological analysis. The proposed method is based on a fully convolutional semantic segmentation scheme that provides the pixel-wise classification of important blastocyst components that help to automatically check the morphologies of these elements. The proposed SSS-Net uses the sprint convolutional block (SCB), which uses asymmetric kernel convolutions in combination with depth-wise separable convolutions to reduce the overall cost of the network. SSS-Net is a shallow architecture with dense feature aggregation, which helps in better segmentation. The proposed SSS-Net consumes a smaller number of trainable parameters (4.04 million) compared to state-of-the-art methods. The SSS-Net was evaluated using a publicly available human blastocyst image dataset for component segmentation. The experimental results confirm that our proposal provides promising segmentation performance with a Jaccard Index of 82.88%, 77.40%, 88.39%, 84.94%, and 96.03% for ZP, TE, BL, ICM, and background, with residual connectivity, respectively. It is also provides a Jaccard Index of 84.51%, 78.15%, 88.68%, 84.50%, and 95.82% for ZP, TE, BL, ICM, and background, with dense connectivity, respectively. The proposed SSS-Net is providing a mean Jaccard Index (Mean JI) of 85.93% and 86.34% with residual and dense connectivity, respectively; this shows effective segmentation of blastocyst components for embryological analysis.
RESUMO
BACKGROUND: Early recognition of prostheses before reoperation can reduce perioperative morbidity and mortality. Because of the intricacy of the shoulder biomechanics, accurate classification of implant models before surgery is fundamental for planning the correct medical procedure and setting apparatus for personalized medicine. Expert surgeons usually use X-ray images of prostheses to set the patient-specific apparatus. However, this subjective method is time-consuming and prone to errors. METHOD: As an alternative, artificial intelligence has played a vital role in orthopedic surgery and clinical decision-making for accurate prosthesis placement. In this study, three different deep learning-based frameworks are proposed to identify different types of shoulder implants in X-ray scans. We mainly propose an efficient ensemble network called the Inception Mobile Fully-Connected Convolutional Network (IMFC-Net), which is comprised of our two designed convolutional neural networks and a classifier. To evaluate the performance of the IMFC-Net and state-of-the-art models, experiments were performed with a public data set of 597 de-identified patients (597 shoulder implants). Moreover, to demonstrate the generalizability of IMFC-Net, experiments were performed with two augmentation techniques and without augmentation, in which our model ranked first, with a considerable difference from the comparison models. A gradient-weighted class activation map technique was also used to find distinct implant characteristics needed for IMFC-Net classification decisions. RESULTS: The results confirmed that the proposed IMFC-Net model yielded an average accuracy of 89.09%, a precision rate of 89.54%, a recall rate of 86.57%, and an F1.score of 87.94%, which were higher than those of the comparison models. CONCLUSION: The proposed model is efficient and can minimize the revision complexities of implants.
RESUMO
Hybrids based on carbon nanotubes (CNTs) and graphene nanoribbons (GNRs) are expected to have synergistic effects for various applications. Herein, we demonstrate a simple one-pot synthesis of a CNT/GNR hybrid material by adjusting the oxidation and unzipping conditions of multi-walled CNTs (MWNTs). The MWNT/graphene oxide nanoribbon (GONR) hybrid was dispersed in various solvents, particularly showing the hybrid hydrogel phase in water at a concentration of 40 mg mL-1. The MWNT/GONR hydrogel exhibited shear-thinning behavior, which can be beneficial for coating a large-area MWNT/GONR layer onto a polymeric porous support by using a scalable slot-die coater. The MWNT/GONR membrane exhibited an outstanding nanofiltration performance, with a molecular weight cutoff of 300 Da and a dye/salt diafiltration performance with a separation factor of 1000 and a water flux of 367.8 LMH, far surpassing the upper bound of diafiltration performance of the existing membranes.
RESUMO
OBJECTIVE: The findings from epidemiological studies addressing the association between adiposity and the risk of colorectal adenomatous polyps are inconsistent. We performed a meta-analysis of epidemiological studies including cross-sectional, case-control, and cohort studies. METHODS: We searched PubMed and EMBASE in June, 2010. All searched articles were reviewed and selected independently by two evaluators according to pre-determined selection criteria. RESULTS: We included 25 studies (nine cross-sectional studies, eleven case-control studies, and five prospective cohort studies) that comprised a total of 300,671 participants and 20,903 cases in the final analysis. When all studies were pooled, the odds ratio (OR) or relative risk (RR) of adiposity and abdominal adiposity for colorectal adenomatous polyp risk was 1.43 (95% confidence interval (CI) 1.23-1.67; n = 22) and 1.42 (95% CI 1.30-1.56; n = 12), respectively. Similarly, an increased risk of colorectal adenomatous polyps was observed in most of the subgroup meta-analyses. CONCLUSIONS: Overall, we found that adiposity and abdominal adiposity significantly increased the risk of colorectal adenomatous polyps in a meta-analysis of epidemiological studies.
Assuntos
Pólipos Adenomatosos/etiologia , Adiposidade/fisiologia , Neoplasias Colorretais/etiologia , Pólipos Adenomatosos/epidemiologia , Adulto , Idoso , Algoritmos , Estudos de Casos e Controles , Estudos de Coortes , Neoplasias Colorretais/epidemiologia , Estudos Transversais , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Fatores de RiscoRESUMO
Pulmonary function impairment has a connection with abdominal obesity, type 2 diabetes, and insulin resistance. Sex differences in lifestyle factors, and pulmonary structure and function may affect pulmonary function in different manners. This study focused on sex differences in the relationship of MetS and its component with pulmonary function. Among 2,614 Korean adults (1,059 men; 1,555 women), pulmonary function was measured by the percentage of predicted forced vital capacity (FVC (%)) and a ratio between forced expiratory volume in 1 second (FEV(1))/FVC. FVC (%) and FEV(1)/FVC were compared according to the presence of MetS and its components. Multiple linear regression analysis was conducted to assess the association between FVC (%), FEV(1)/FVC and clinical variables. We found sex differences in the relationship of MetS and its components with pulmonary function. FVC (%) was significantly lower in subjects with MetS than in those without MetS in both men and women, and FEV(1)/FVC was lower in subjects with MetS only in women. Among components of MetS, waist circumference, blood pressure and fasting plasma glucose, and HDL-cholesterol were independently related to FVC (%) in men, whereas waist circumference was significantly associated with FVC (%) in women. Blood pressure was found to be an independent factor of FEV(1)/FVC in men, whereas blood pressure, fasting plasma glucose and HDL-cholesterol independently determined FEV(1)/FVC in women. These findings suggest that sex-specific association between MetS and lung function measures should be considered in clinical practice.
Assuntos
Pulmão/fisiopatologia , Síndrome Metabólica/fisiopatologia , Adulto , Idoso , Povo Asiático , Glicemia , Pressão Sanguínea , HDL-Colesterol/sangue , Estudos Transversais , Feminino , Volume Expiratório Forçado/fisiologia , Humanos , Masculino , Pessoa de Meia-Idade , Inquéritos Nutricionais , República da Coreia/epidemiologia , Caracteres Sexuais , Capacidade Vital/fisiologia , Circunferência da CinturaRESUMO
Social media has become an ideal platform for the propagation of rumors, fake news, and misinformation. Rumors on social media not only mislead online users but also affect the real world immensely. Thus, detecting the rumors and preventing their spread became an essential task. Some of the recent deep learning-based rumor detection methods, such as Bi-Directional Graph Convolutional Networks (Bi-GCN), represent rumor using the completed stage of the rumor diffusion and try to learn the structural information from it. However, these methods are limited to represent rumor propagation as a static graph, which isn't optimal for capturing the dynamic information of the rumors. In this study, we propose novel graph convolutional networks with attention mechanisms, named Dynamic GCN, for rumor detection. We first represent rumor posts with their responsive posts as dynamic graphs. The temporal information is used to generate a sequence of graph snapshots. The representation learning on graph snapshots with attention mechanism captures both structural and temporal information of rumor spreads. The conducted experiments on three real-world datasets demonstrate the superiority of Dynamic GCN over the state-of-the-art methods in the rumor detection task.
Assuntos
Comunicação , Desinformação , Atividades Humanas/psicologia , Disseminação de Informação/métodos , Redes Neurais de Computação , Reconhecimento Psicológico/fisiologia , Mídias Sociais/normas , Humanos , Modelos TeóricosRESUMO
Retinal blood vessels are considered valuable biomarkers for the detection of diabetic retinopathy, hypertensive retinopathy, and other retinal disorders. Ophthalmologists analyze retinal vasculature by manual segmentation, which is a tedious task. Numerous studies have focused on automatic retinal vasculature segmentation using different methods for ophthalmic disease analysis. However, most of these methods are computationally expensive and lack robustness. This paper proposes two new shallow deep learning architectures: dual-stream fusion network (DSF-Net) and dual-stream aggregation network (DSA-Net) to accurately detect retinal vasculature. The proposed method uses semantic segmentation in raw color fundus images for the screening of diabetic and hypertensive retinopathies. The proposed method's performance is assessed using three publicly available fundus image datasets: Digital Retinal Images for Vessel Extraction (DRIVE), Structured Analysis of Retina (STARE), and Children Heart Health Study in England Database (CHASE-DB1). The experimental results revealed that the proposed method provided superior segmentation performance with accuracy (Acc), sensitivity (SE), specificity (SP), and area under the curve (AUC) of 96.93%, 82.68%, 98.30%, and 98.42% for DRIVE, 97.25%, 82.22%, 98.38%, and 98.15% for CHASE-DB1, and 97.00%, 86.07%, 98.00%, and 98.65% for STARE datasets, respectively. The experimental results also show that the proposed DSA-Net provides higher SE compared to the existing approaches. It means that the proposed method detected the minor vessels and provided the least false negatives, which is extremely important for diagnosis. The proposed method provides an automatic and accurate segmentation mask that can be used to highlight the vessel pixels. This detected vasculature can be utilized to compute the ratio between the vessel and the non-vessel pixels and distinguish between diabetic and hypertensive retinopathies, and morphology can be analyzed for related retinal disorders.
RESUMO
[This corrects the article .].