Domain Adaptation with Contrastive Simultaneous Multi-Loss Training for Hand Gesture Recognition.

Hand gesture recognition from images is a critical task with various real-world applications, particularly in the field of human-robot interaction. Industrial environments, where non-verbal communication is preferred, are significant areas of application for gesture recognition. However, these environments are often unstructured and noisy, with complex and dynamic backgrounds, making accurate hand segmentation a challenging task. Currently, most solutions employ heavy preprocessing to segment the hand, followed by the application of deep learning models to classify the gestures. To address this challenge and develop a more robust and generalizable classification model, we propose a new form of domain adaptation using multi-loss training and contrastive learning. Our approach is particularly relevant in industrial collaborative scenarios, where hand segmentation is difficult and context-dependent. In this paper, we present an innovative solution that further challenges the existing approach by testing the model on an entirely unrelated dataset with different users. We use a dataset for training and validation and demonstrate that contrastive learning techniques in simultaneous multi-loss functions provide superior performance in hand gesture recognition compared to conventional approaches in similar conditions.

Sediment Microbial Diversity of Three Deep-Sea Hydrothermal Vents Southwest of the Azores.

Menez Gwen, Lucky Strike and Rainbow are the three most visited and well-known deep-sea hydrothermal vent fields in the Azores region, located in the Mid-Atlantic Ridge. Their distinct geological and ecological features allow them to support a diversity of vent communities, which are largely dependent on Bacteria and Archaea capable of anaerobic or microaerophilic metabolism. These communities play important ecological roles through chemoautotrophy, feeding and in establishing symbiotic associations. However, the occurrence and distribution of these microbes remain poorly understood, especially in deep-sea sediments. In this study, we provide for the first time a comparative survey of the sediment-associated microbial communities from these three neighbouring vent fields. Sediment samples collected in the Menez Gwen, Lucky Strike and Rainbow vent fields showed significant differences in trace-metal concentrations and associated microbiomes. The taxonomic profiles of bacterial, archaeal and eukaryotic representatives were assessed by rRNA gene-tag pyrosequencing, identified anaerobic methanogens and microaerobic Epsilonproteobacteria, particularly at the Menez Gwen site, suggesting sediment communities potentially enriched in sub-seafloor microbes rather than from pelagic microbial taxa. Cosmopolitan OTUs were also detected mostly at Lucky Strike and Rainbow sites and affiliated with the bacterial clades JTB255, Sh765B-TzT-29, Rhodospirillaceae and OCS155 marine group and with the archaeal Marine Group I. Some variations in the community composition along the sediment depth were revealed. Elemental contents and hydrothermal influence are suggested as being reflected in the composition of the microbial assemblages in the sediments of the three vent fields. Altogether, these findings represent valuable information for the understanding of the microbial distribution and potential ecological roles in deep-sea hydrothermal fields.

Microbial diversity in deep-sea sediments from the Menez Gwen hydrothermal vent system of the Mid-Atlantic Ridge.

Deep-sea hydrothermal sediments are known to support remarkably diverse microbial consortia. Cultureindependent sequence-based technologies have extensively been used to disclose the associated microbial diversity as most of the microorganisms inhabiting these ecosystems remain uncultured. Here we provide the first description of the microbial community diversity found on sediments from Menez Gwen vent system. We compared hydrothermally influenced sediments, retrieved from an active vent chimney at 812 m depth, with non-hydrothermally influenced sediments, from a 1400 m depth bathyal plain. Considering the enriched methane and sulfur composition of Menez Gwen vent fluids, and the sediment physicochemical properties in each sampled area, we hypothesized that the site-associated microbes would be different. To address this question, taxonomic profiles of bacterial, archaeal and micro-eukaryotic representatives were studied by rRNA gene tag pyrosequencing. Communities were shown to be significantly different and segregated by sediment geographical area. Specific mesophilic, thermophilic and hyperthermophilic archaeal (e.g., Archaeoglobus, ANME-1) and bacterial (e.g., Caldithrix, Thermodesulfobacteria) taxa were highly abundant near the vent chimney. In contrast, bathyal-associated members affiliated to more ubiquitous phylogroups from deep-ocean sediments (e.g., Thaumarchaeota MGI, Gamma- and Alphaproteobacteria). This study provides a broader picture of the biological diversity and microbial biogeography, and represents a preliminary approach to the microbial ecology associated with the deep-sea sediments from the Menez Gwen hydrothermal vent field.

Wine fermentation microbiome: a landscape from different Portuguese wine appellations.

Grapes and wine musts harbor a complex microbiome, which plays a crucial role in wine fermentation as it impacts on wine flavour and, consequently, on its final quality and value. Unveiling the microbiome and its dynamics, and understanding the ecological factors that explain such biodiversity, has been a challenge to oenology. In this work, we tackle this using a metagenomics approach to describe the natural microbial communities, both fungal and bacterial microorganisms, associated with spontaneous wine fermentations. For this, the wine microbiome, from six Portuguese wine appellations, was fully characterized as regards to three stages of fermentation - Initial Musts (IM), and Start and End of alcoholic fermentations (SF and EF, respectively). The wine fermentation process revealed a higher impact on fungal populations when compared with bacterial communities, and the fermentation evolution clearly caused a loss of the environmental microorganisms. Furthermore, significant differences (p < 0.05) were found in the fungal populations between IM, SF, and EF, and in the bacterial population between IM and SF. Fungal communities were characterized by either the presence of environmental microorganisms and phytopathogens in the IM, or yeasts associated with alcoholic fermentations in wine must samples as Saccharomyces and non-Saccharomyces yeasts (as Lachancea, Metschnikowia, Hanseniaspora, Hyphopichia, Sporothrix, Candida, and Schizosaccharomyces). Among bacterial communities, the most abundant family was Enterobacteriaceae; though families of species associated with the production of lactic acid (Lactobacillaceae, Leuconostocaceae) and acetic acid (Acetobacteriaceae) were also detected. Interestingly, a biogeographical correlation for both fungal and bacterial communities was identified between wine appellations at IM suggesting that each wine region contains specific and embedded microbial communities which may contribute to the uniqueness of regional wines.

Unravelling the diversity of grapevine microbiome.

Vitis vinifera is one of the most widely cultivated fruit crops with a great economic impact on the global industry. As a plant, it is naturally colonised by a wide variety of both prokaryotic and eukaryotic microorganisms that interact with grapevine, having either beneficial or phytopathogenic effects, who play a major role in fruit yield, grape quality and, ultimately, in the evolution of grape fermentation and wine production. Therefore, the objective of this study was to extensively characterize the natural microbiome of grapevine. Considering that the majority of microorganisms are uncultivable, we have deeply studied the microflora of grapevine leaves using massive parallel rDNA sequencing, along its vegetative cycle. Among eukaryotic population the most abundant microorganisms belonged to the early diverging fungi lineages and Ascomycota phylum, whereas the Basidiomycota were the least abundant. Regarding prokaryotes, a high diversity of Proteobacteria, Firmicutes and Actinobacteria was unveiled. Indeed, the microbial communities present in the vineyard during its vegetative cycle were shown to be highly structured and dynamic. In all cases, the major abundant microorganisms were the yeast-like fungus Aureobasidium and the prokaryotic Enterobacteriaceae. Herein, we report the first complete microbiome landscape of the vineyard, through a metagenomic approach, and highlight the analysis of the microbial interactions within the vineyard and its importance for the equilibrium of the microecosystem of grapevines.

Assessment of the geographic origins of pinewood nematode isolates via single nucleotide polymorphism in effector genes.

The pinewood nematode, Bursaphelenchus xylophilus, is native to North America but it only causes damaging pine wilt disease in those regions of the world where it has been introduced. The accurate detection of the species and its dispersal routes are thus essential to define effective control measures. The main goals of this study were to analyse the genetic diversity among B. xylophilus isolates from different geographic locations and identify single nucleotide polymorphism (SNPs) markers for geographic origin, through a comparative transcriptomic approach. The transcriptomes of seven B. xylophilus isolates, from Continental Portugal (4), China (1), Japan (1) and USA (1), were sequenced in the next generation platform Roche 454. Analysis of effector gene transcripts revealed inter-isolate nucleotide diversity that was validated by Sanger sequencing in the genomic DNA of the seven isolates and eight additional isolates from different geographic locations: Madeira Island (2), China (1), USA (1), Japan (2) and South Korea (2). The analysis identified 136 polymorphic positions in 10 effector transcripts. Pairwise comparison of the 136 SNPs through Neighbor-Joining and the Maximum Likelihood methods and 5-mer frequency analysis with the alignment-independent bilinear multivariate modelling approach correlated the SNPs with the isolates geographic origin. Furthermore, the SNP analysis indicated a closer proximity of the Portuguese isolates to the Korean and Chinese isolates than to the Japanese or American isolates. Each geographic cluster carried exclusive alleles that can be used as SNP markers for B. xylophilus isolate identification.