Semantic Segmentation for People Detection on Beach Images

Nosso trabalho compara um conjunto de redes de segmentação semântica aplicados na detecção de pessoas em imagens de praia, como parte de um sistema de rastreamento automático para evitar que banhistas ultrapassem a região segura do mar. Em nossa análise, comparamos as redes de segmentação U-net, X-net, Linknet, e Unet++ usando os backbones prétreinados VGG-16 e VGG-19. Nós propomos nossa própria base de imagens, composta de 300 imagens. Os modelos foram avaliados utilizando a métrica F-score. Nossos resultados mostraram que a Linknet obteve o melhor valor de F-score, com 90.89%, enquanto a Linknet foi mais rápida que as outras redes, sem diferença estatística significativa.

Features of systemic cytokine production in Behcet's disease associated with uveitis without ocular lesions

Behcet's disease (BD) is a systemic disease underlyed by chronic vasculitis. Hyperactivity of innate and adaptive immunity plays important role in its pathogenesis. Uveitis occurs in 30-70% of the patients, often recurring and reducing visual function. The objective of our work was to study the features of systemic production of immune mediators in BD patients, depending on presence and activity of uveitis. 116 BD patients were divided into 3 groups: (1) 41 patients with active uveitis (UA), (2) 64 subjects with uveitis remission (UR), (3) 11 uveitis-free BD patients (WU). Control group (CG) comprised 34 conditionally healthy people. Detection rate (%) and contents (pg/ml) were measured for IL-1β IL-2, IL-4, IL-5, IL-6, IL-12p70, IL-13, IL-18, IFNγ, CCL2/MCP-1, CCL3/MIP-1α, CCL4/MIP-1β, CCL5/RANTES, CCL11/Eotaxin, СXCL1/GRO-α, CXCL8/IL-8, CXCL10/IP-10, CXCL12/SDF-1α, GM-CSF, TNFα in blood serum by means of multiplex analysis using MAGPIX analyzer (Luminex Corp., USA), Procarta Plex “Human Th1/Th2&Chemokine Panel 20 plex” kits (Bioscience, Austria). TGF-P1, TGF-P2 levels were assayed by ELISA-test (“Vfector-Best”). All the BD patients showed high detection rates of CXCL1/GRO-α (but not its level) in comparison with CG. Detection rate and levels of IL-6, IL-8 were increased in 1st and 2nd BD groups, compared to CG. In UR, unlike UA and WU groups, IL-4 was detected more often than in CG. WU patients showed increased detection rate of only CXCL1/GRO -α. When compared with UA, WU patients had lower serum concentrations of IFNγ, MCP-1, IP-10, MIP-1a, SDF-1α, TGF-β1; UR patients also showed decreased serum levels of IL-18, Eotaxin, GRO-α, RANTES, TGF-β2. Our results indicate the importance of angiogenic and proinflammatory chemokines and cytokines in pathogenesis of BD uveitis, as well as imbalanced production of various immunomediators. Higher detection rates and levels of IL-6 and IL-8 in UA and UR patients may result from weak persistent intraocular inflammation, even upon relief of clinical symptoms, thus, probably, requiring therapeutic correction.

IFND: a benchmark dataset for fake news detection

Spotting fake news is a critical problem nowadays. Social media are responsible for propagating fake news. Fake news propagated over digital platforms generates confusion as well as induce biased perspectives in people. Detection of misinformation over the digital platform is essential to mitigate its adverse impact. Many approaches have been implemented in recent years. Despite the productive work, fake news identification poses many challenges due to the lack of a comprehensive publicly available benchmark dataset. There is no large-scale dataset that consists of Indian news only. So, this paper presents IFND (Indian fake news dataset) dataset. The dataset consists of both text and images. The majority of the content in the dataset is about events from the year 2013 to the year 2021. Dataset content is scrapped using the Parsehub tool. To increase the size of the fake news in the dataset, an intelligent augmentation algorithm is used. An intelligent augmentation algorithm generates meaningful fake news statements. The latent Dirichlet allocation (LDA) technique is employed for topic modelling to assign the categories to news statements. Various machine learning and deep-learning classifiers are implemented on text and image modality to observe the proposed IFND dataset's performance. A multi-modal approach is also proposed, which considers both textual and visual features for fake news detection. The proposed IFND dataset achieved satisfactory results. This study affirms that the accessibility of such a huge dataset can actuate research in this laborious exploration issue and lead to better prediction models.