Search Results (306)

3D visual grounding is a fundamental task for human–machine interaction, aiming to localize specific objects in complex 3D point clouds based on natural language descriptions. Despite recent advancements, existing Transformer-based architectures often rely on absolute position embeddings and heuristic query initialization, which lack the capacity to capture fine-grained relative spatial dependencies and fail to effectively filter out scene clutter. In this paper, we propose SESQ, a novel framework that synergizes Spatially Aware Encoding and Semantically Guided Querying for 3D grounding. Our approach introduces two key innovations. First, we propose the Rotary Spatially Aware Encoder (RSAE), which incorporates Rotary Position Embeddings (RoPE) into the self-attention layers. By transforming 3D coordinates into a rotary representation, RSAE enables the model to inherently capture relative spatial distances and maintains geometric consistency throughout the encoding stage. Second, a Semantic Query Initialization (SQI) module is designed to initialize object queries by explicitly computing the cross-modal similarity between textual embeddings and visual point cloud features. By replacing traditional heuristic-based selection with semantic-aware alignment, SQI ensures that the decoding process originates from contextually relevant object candidates, significantly reducing the impact of task-irrelevant distractors. Extensive experiments on ScanRefer and ReferIt3D (Nr3D/Sr3D) benchmarks demonstrate the effectiveness of our framework. Compared to the baseline EDA, our method achieves a significant performance gain of 2.68% in overall Acc@0.5 on ScanRefer, a 4.9% improvement on the challenging Nr3D “Hard” subset, and a 1.1% increase in overall Acc@0.25 on Sr3D. Full article

Background: Interoception has been proposed as a key mechanism underlying impulsive behaviours, including maladaptive eating. However, the brain mechanisms supporting the interaction between interoception and impulsivity across different reward types remain unclear. This study investigated whether modulating the right insula and the dorsal anterior cingulate cortex (dACC) using high-definition transcranial direct current stimulation (HD-tDCS) could affect interoceptive accuracy and impulsive decision-making. Methods: Model-based HD-tDCS montages were defined to target the right insula and dACC. Two behavioural paradigms were administered: (i) the heartbeat detection task (HBD) to assess interoceptive accuracy and (ii) two versions of the delay discounting (DD) task with food and monetary rewards to measure impulsivity. Heart rate variability (HRV) was recorded as an index of autonomic activity. HD-tDCS was delivered online during the HBD, while DD tasks were completed offline. Twenty-four participants took part in four sessions in a within-subject design: baseline DD tasks, anodal HD-tDCS targeting the insula, dACC, or sham stimulation. Results: Stimulation of both the insula and dACC reduced participants’ ability to detect synchronous heartbeat while improving accuracy in exteroceptive trials. Discounting rates significantly increased following insula stimulation. Moreover, HD-tDCS effects on DD performance varied depending on reward type. Conclusions: These findings suggest differential contributions of the dACC and insula in interoceptive and exteroceptive processing and support the effect of HD-tDCS combined with interoceptive tasks to modulate impulsive decision-making. Reward-specific effects highlight the importance of stimulus type when designing interventions for impulsive eating behaviours. Full article

Accurate quantification of spatial distances between fluorescent signals in multi-channel 3D microscopy is essential for understanding genomic organization and gene regulation. However, chromatic aberration introduces systematic spatial offsets between channels that significantly bias distance measurements, particularly at short genomic distances. We present FISH-Dist, an automated computational pipeline for quantitative distance measurements in 3D fluorescence in situ hybridization (FISH) experiments acquired on standard confocal microscopes. Our method combines deep learning-based spot segmentation, 3D Gaussian fitting for sub-pixel localization, and two complementary chromatic aberration correction approaches: affine (ACC) and linear (LCC). We validated the pipeline by measuring the lengths of DNA origami nanorulers and systematically evaluated FISH probe design parameters, including probe spacing, density, and target sequence length. FISH-Dist achieves sub-pixel accuracy in signal detection and substantially reduces inter-channel distance measurement errors. This enables a reproducible quantification of spatial relationships in 3D FISH datasets. Unlike existing tools optimized for long-range chromosomal interactions or requiring super-resolution microscopy, FISH-Dist specifically addresses the technical challenges of standard confocal imaging at short genomic distances, where chromatic aberration has a proportionally greater impact on measurement accuracy. Full article

Obesity is a complex metabolic disorder characterized by excessive accumulation of adipose tissue, resulting from an imbalance between energy intake and expenditure. It is associated with an increased risk of chronic diseases such as type 2 diabetes, cardiovascular disease, and cancer. Kefiran is a water-soluble exopolysaccharide produced by lactic acid bacteria, Lactobacillus kefiranofaciens, in kefir grains, composed primarily of glucose and galactose. It has garnered scientific interest due to its antioxidant, anti-inflammatory, and antimicrobial properties. Rice Kefiran (RK) is a functional food made with culturing L. kefiranofaciens in a medium containing rice. It is standardized to contain at least 5 mg/g of kefiran. This study investigated the anti-obesity effect of RK on a high-fat diet (HFD)-induced obese mouse model. HFD-fed mice exhibited marked increases in body weight gain (10.3 g vs. 2.0 g in controls) and adipose tissue mass (2.4 g vs. 0.4 g in controls). RK administration significantly attenuated weight gain to 8.3 g and 6.0 g at doses of 10 and 50 mg/kg, respectively, and reduced adipose tissue mass to 2.2 g (RK10) and 1.7 g (RK50). Oral glucose tolerance testing revealed impaired glucose clearance in HFD-fed mice, with blood glucose levels of 403.5 mg/dL at 15 min and 314.6 mg/dL at 120 min, compared with 348.8 mg/dL and 232.2 mg/dL in controls. RK treatment improved glucose tolerance, particularly at 50 mg/kg, reducing glucose levels to 359.0 mg/dL at 15 min and 263.8 mg/dL at 120 min. Biochemical analyses demonstrated that RK significantly reduced serum total cholesterol (213.6 mg/dL in HFD vs. 178.0 and 184.0 mg/dL in RK10 and RK50), triglycerides (379.0 mg/dL in HFD vs. 228.8 and 234.6 mg/dL), and non-esterified fatty acids (0.89 mEq/mL in HFD vs. 0.54 and 0.35 mEq/mL), while phospholipid levels remained unchanged. Furthermore, RK increased serum nicotinamide phosphoribosyltransferase (NAMPT) levels from 15.8 ng/mL in HFD-fed mice to 30.0 and 50.0 ng/mL in the RK10 and RK50 groups, respectively, and restored hepatic NAD⁺/NADH ratios toward control levels (1.78 µmol/L in HFD vs. 1.90 µmol/L and 2.07 µmol/L in RK10 and RK50). Gene expression analysis showed that RK increased Nampt mRNA expression and decreased the mRNA expression of adipogenic and lipogenic genes, including Srebp-1c, Acc-1, and Fas. These findings suggest that RK may ameliorate obesity-related metabolic disturbances and its associated metabolic dysfunctions by modulating lipid metabolism, glucose tolerance, and NAD⁺ biosynthesis pathways. Full article

From Palaeolithic ornaments to modern biomimetics, the use of nacre and shells has evolved. Initially utilised for jewellery and tools, they now inspire the development of advanced materials. This paper reviews the current knowledge on nacre’s composition, focusing on the highly regulated biomineralisation process wherein amorphous calcium carbonate (ACC) transforms into crystalline aragonite. It examines the important role of the organic matrix (specifically soluble, insoluble, and acidic proteins) in controlling crystal nucleation, growth, and polymorph selection. Scientists study natural nacre formation to create nacre-inspired composites for various applications. Charles Hatchett’s in 1799 shell categorisation, Sorby and Sowerby’s 19th-century microscopy, Taylor, Beedham, Bøggild, and Currey’s mid-20th-century research on bivalve structures, and mechanical property investigations in the 1970s are some of the major developments. The hierarchical structure, cooperative plastic deformation, surface asperities, organic–inorganic interactions, and interphase in such complex composite materials give rise to impressive mechanical properties. In the early 2000s, with the emergence of biomimetics, inspired by nacre, several macroscopic structural materials with uniform micro- and nanoscale architectures have been synthesised in recent decades, and their mechanical properties and potential applications have been explored. Modern nacre-inspired fabrication utilises 3D printing for precision, freeze casting for sustainability, and mineralisation for scalability. Techniques like layer-by-layer assembly and nanomaterial integration enhance mechanical performance through advanced interfacial engineering. Full article

Background: Viburnum (Adoxaceae) is a species-rich woody genus whose taxonomy is complicated by morphological convergence and hybridization. Methods: We assembled complete plastomes of eight species representing five sections and analyzed their structural variation, sequence divergence, and molecular evolution. Results: All plastomes displayed the conserved quadripartite structure typical of angiosperms, with limited size variation attributable primarily to intergenic spacer-length polymorphisms. Sequence divergence was unevenly distributed, with single-copy regions exhibiting substantially higher nucleotide diversity than inverted repeat regions. We identified multiple hypervariable intergenic spacers such as the region trnK-UUU–rps16, suitable as molecular markers for population genetics and species identification. Selection pressure analysis revealed that while most protein-coding genes evolved under strong purifying selection, three genes involved in fatty acid biosynthesis and protein import—accD, ycf1, and ycf2—showed significantly relaxed constraints, suggesting ongoing functional divergence. Phylogenetic analysis recovered well-supported relationships consistent with previous classifications, while clarifying the positions of Viburnum amplificatum and Viburnum tinus. Conclusions: These findings provide molecular resources for Viburnum systematics and offer insights into the evolutionary dynamics of plastome genes with non-photosynthetic functions. Full article

To maintain the integrity of the outer membrane of Gram-negative bacteria, such as Escherichia coli, the levels of two essential components, phospholipids (PL) and lipopolysaccharide (LPS), are tightly regulated, although the underlying molecular mechanisms are unclear. E. coli synthesizes three main PLs, including essential phosphatidylethanolamine and phosphatidylglycerol and nonessential cardiolipin (CL). We showed that CL synthesis is conditionally essential in ΔlapD bacteria. Using this synthetic lethal phenotype, we isolated suppressors that rescued growth at elevated temperatures. We showed that loss-of-function mutations in cdsA encoding CDP-diglyceride synthetase, and pgsA, which encodes phosphatidylglycerophosphate synthase, bypass this lethality. Such mutations reduce the relative abundance of acidic phospholipids, which are otherwise elevated in Δ(lapD clsA) bacteria, and increase the amounts of cis-vaccenic acid without altering amounts of LpxC mediating the first committed step in LPS biosynthesis. Interestingly, overexpression of genes, including accC and glnB, whose products can inhibit fatty acid/PL synthesis, overcame the lethality of Δ(lapD clsA) bacteria. We demonstrated that PgsA co-purifies with LapB, which regulates LpxC stability and acts as a hub for proteins involved in PL and LPS biosynthesis, including LapD. Overall, our results reveal that LapD is positioned at the regulatory nexus between LPS assembly and fatty acid/PL synthesis. Full article

The identification of community structure is pivotal for understanding the functional characteristics of complex networks. To address the limitations of most existing community detection algorithms, which often require predefining the number of communities and lack robustness, this paper proposes a novel community detection algorithm named D-means (K-means community detection algorithm based on density peaks). This algorithm integrates the concept of density peak clustering with K-means spectral clustering, employing Chebyshev’s inequality to automatically determine the number of community centers, thereby enabling unsupervised identification of community quantities. By designing a multi-dimensional evaluation framework, the comparative experiments were conducted on LFR benchmark networks (Lancichinetti-Fortunato-Radicchi benchmark networks) and real-world social network datasets. The results demonstrate that the D-means algorithm outperforms traditional algorithms in terms of ACC (accuracy), ARI (adjusted rand index), and NMI (normalized mutual information) metrics, while also achieving improvements in runtime efficiency, showcasing strong robustness. Finally, the D-means algorithm was applied to the public transportation network of Urumqi. Empirical analysis identified 12 functionally significant transportation communities, providing theoretical support for urban rail transit optimization and commercial facility layout planning. Full article

The flowers of Panax ginseng C. A. Mey. (PG) and Panax notoginseng (Burkill) F. H. Chen ex C. H. Chow (PN) are morphologically indistinguishable after drying, leading to prevalent adulteration that compromises product quality and consumer safety. To address this issue, we developed a rapid, closed-tube molecular authentication method based on high-resolution melting (HRM) analysis. Species-specific primer pairs were designed to target the conserved ITS and rbcL-accD regions, with PNG-2 selected as the optimal candidate owing to its stable genotyping performance and moderate GC content. Our results established GC content, rather than amplicon length, as the primary determinant of the melting temperature (Tm). Notably, the experimentally measured Tm values were consistently 0.7–1.5 °C higher than theoretical predictions, a discrepancy attributable to the stabilizing effect of the saturated fluorescent dye. To ensure maximum diagnostic reliability, the HRM results were cross-validated through a three-tier system comprising ITS2 phylogenetic analysis, agarose gel electrophoresis, and Sanger sequencing. The practical utility and matrix robustness of the assay were further verified using a diversified validation cohort of 30 commercial samples, including 24 floral batches and 6 root-derived products (root slices and ultramicro powders). The HRM profiles demonstrated 100% concordance with DNA barcoding results, effectively identifying mislabeled products across different botanical matrices and processing forms. This methodology, which can be completed within 3 h, provides a significantly more cost-effective and rapid alternative to traditional sequencing-based methods for large-scale market surveillance and industrial quality control. Full article

High-precision temperature prediction is crucial for dealing with extreme weather events under the background of global warming. However, due to the limitations of computing resources, numerical weather prediction models are difficult to directly provide high spatio-temporal resolution data that meets the specific application requirements of a certain region. This problem is particularly prominent in areas with complex terrain. The use of remote sensing data, especially high-resolution terrain data, provides key information for understanding and simulating the interaction between land and atmosphere in complex terrain, making the integration of remote sensing and NWP outputs to achieve high-precision meteorological element downscaling a core challenge. Aiming at the challenge of temperature scaling in complex terrain areas of Southwest China, this paper proposes a novel deep learning model—Terrain Adaptive U-Net Transformer (TAUT). This model takes the encoder–decoder structure of U-Net as the skeleton, deeply integrates the global attention mechanism of Swin Transformer and the local spatiotemporal feature extraction ability of three-dimensional convolution, and innovatively introduces the multi-branch terrain adaptive module (MBTA). The adaptive integration of terrain remote sensing data with various meteorological data, such as temperature fields and wind fields, has been achieved. Eventually, in the complex terrain area of Southwest China, a spatio-temporal high-resolution downscaling of 2 m temperature was realized (from 0.1° in space to 0.01°, and from 3 h intervals to 1 h intervals in time). The experimental results show that within the 48 h downscaling window period, the TAUT model outperforms the comparison models such as bilinear interpolation, SRCNN, U-Net, and EDVR in all evaluation metrics (MAE, RMSE, COR, ACC, PSNR, SSIM). The systematic ablation experiment verified the independent contributions and synergistic effects of the Swin Transformer module, the 3D convolution module, and the MBTA module in improving the performance of each model. In addition, the regional terrain verification shows that this model demonstrates good adaptability and stability under different terrain types (mountains, plateaus, basins). Especially in cases of high-temperature extreme weather, it can more precisely restore the temperature distribution details and spatial textures affected by the terrain, verifying the significant impact of terrain remote sensing data on the accuracy of temperature downscaling. The core contribution of this study lies in the successful construction of a hybrid architecture that can jointly leverage the local feature extraction advantages of CNN and the global context modeling capabilities of Transformer, and effectively integrate key terrain remote sensing data through dedicated modules. The TAUT model offers an effective deep learning solution for precise temperature prediction in complex terrain areas and also provides a referential framework for the integration of remote sensing data and numerical model data in deep learning models. Full article

Orbea is a morphologically diverse lineage within the subtribe Stapeliinae, yet plastome evolution in Arabian taxa remains insufficiently characterized. This study reports the first complete chloroplast genomes of Orbea sprengeri subsp. commutata and O. wissmannii var. eremastrum and investigates plastome structure, sequence variability, and phylogenetic relationships across tribe Ceropegieae. Chloroplast genomes were assembled, annotated, and compared with 13 published plastomes representing major Ceropegieae lineages. Both Arabian plastomes displayed the typical quadripartite structure and identical gene content of 114 unique genes, including 80 protein-coding genes, 30 transfer RNA genes, and four ribosomal RNA genes. However, O. wissmannii var. eremastrum exhibited pronounced structural divergence, possessing the largest plastome recorded for the tribe (170,054 bp), an 8.9 kb expansion of the inverted repeat regions, and an 8.4 kb inversion spanning the ndhG–ndhF region. Comparative analyses revealed conserved gene order across Ceropegieae but identified six highly variable loci (accD, clpP, ndhF, ycf1, psbM–trnD, and rpl32–trnL) as potential DNA barcodes. Selection pressure analyses indicated strong purifying selection across most genes, with localized adaptive signals in accD, ndhE, ycf1, and ycf2. Phylogenomic reconstruction consistently resolved the two Arabian Orbea taxa as a distinct clade separate from the African O. variegata. This study fills a gap in Ceropegieae plastid genomics and underscores the importance of sequencing additional Orbea species to capture the full extent of genomic variation within this diverse genus. Full article

Brachiaria decumbens is a high-yielding forage grass of major economic value in tropical regions. The root endophytic fungus Piriformospora indica is widely recognized for promoting plant growth and stress tolerance, yet its effects on B. decumbens remain poorly characterized. Here, we profiled root responses to P. indica colonization at 10 days after inoculation (dais; early stage) and 20 dais (late stage) during symbiosis establishment. Colonization was confirmed by phenotypic and physiological assessments, with inoculated plants showing enhanced root growth; colonized roots exhibited higher activities of catalase (CAT), superoxide dismutase (SOD), and peroxidase (POD), along with increased indole-3-acetic acid (IAA) levels, whereas malondialdehyde (MDA), jasmonic acid (JA), and the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) were reduced. Transcriptome and metabolomic profiling identified 1884 and 1077 differentially expressed genes (DEGs) and 2098 and 1509 differentially accumulated metabolites (DAMs) at 10 dais (Pi10d vs. CK10d) and 20 dais (Pi20d vs. CK20d), respectively, and 3355 DEGs and 2314 DAMs between stages (Pi20d vs. Pi10d). Functional enrichment highlighted key pathways related to secondary metabolism, carbohydrate metabolism, and lipid biosynthesis. Differentially expressed transcription factors spanned multiple families, including MYB, AP2/ERF, MADS-box, and bZIP, consistent with broad transcriptional reprogramming during symbiosis establishment. Integrative multi-omics analysis further highlighted phenylpropanoid biosynthesis and α-linolenic acid metabolism as consistently co-enriched pathways, suggesting coordinated shifts in gene expression and metabolite accumulation across colonization stages. Collectively, these results provide a multi-layered resource and a framework for mechanistic dissection of the P. indica–B. decumbens interaction. Full article

Fine-grained visual classification requires recognizing subtle inter-class differences under substantial intra-class variation. Ancient mural recognition poses additional challenges: severe degradation and complex backgrounds introduce noise that obscures discriminative features, limited annotated data restricts model training, and dynasty-specific artistic styles manifest as periodic brushwork patterns and compositional structures that are difficult to capture. Existing spatial-domain methods fail to model the frequency characteristics of textures and the cross-region semantic relationships inherent in mural imagery. To address these limitations, we propose a Vision Transformer (ViT) framework which integrates frequency-domain enhancement, explicit token-relation modeling, adaptive multi-focus inference, and discriminative metric supervision. A Frequency Channel Attention (FreqCA) module applies 2D FFT-based channel gating to emphasize discriminative periodic patterns and textures. A Cross-Token Relation Attention (CTRA) module employs joint global and local gates to strengthen semantically related token interactions across distant regions. An Adaptive Omni-Focus (AOF) block partitions tokens into importance groups for multi-head classification, while Complementary Tokens Integration (CTI) fuses class tokens from multiple transformer layers. Finally, Adaptive Margin Contrastive Center Loss (AMCCL) improves intra-class compactness and inter-class separability with margins adapted to class-center similarities. Experiments on CUB-200-2011, Stanford Dogs, and a Dunhuang mural dataset show accuracies of 91.15%, 94.57%, and 94.27%, outperforming the ACC-ViT baseline by 1.35%, 1.63%, and 2.20%, respectively. Full article

Three-dimensional visual grounding is a core technology that identifies specific objects within complex 3D scenes based on natural language instructions, enhancing human–machine interactions in robotics and augmented reality domains. Traditional approaches have focused on supervised learning, which relies on annotated data; however, zero-shot methodologies are emerging due to the high costs of data construction and limitations in generalization. SeeGround achieves state-of-the-art performance by integrating 2D rendered images and spatial text descriptions. Nevertheless, SeeGround exhibits vulnerabilities in clearly discerning relative depth relationships owing to its implicit depth representations in 2D views. This study proposes the relational depth text (RDT) technique to overcome these limitations, utilizing a Monocular Depth Estimation model to extract depth maps from rendered 2D images and applying the K-Nearest Neighbors algorithm to convert inter-object relative depth relations into natural language descriptions, thereby incorporating them into Vision–Language Model (VLM) prompts. This method distinguishes itself by augmenting spatial reasoning capabilities while preserving SeeGround’s existing pipeline, demonstrating a 3.54% improvement in the Acc@0.25 metric on the Nr3D dataset in a 7B VLM environment that is approximately 10.3 times lighter than the original model, along with a 6.74% increase in Unique cases on the ScanRefer dataset, albeit with a 1.70% decline in Multiple cases. The proposed technique enhances the robustness of grounding through viewpoint anchoring and candidate discrimination in complex query scenarios, and is expected to improve efficiency in practical applications through future multi-view fusion and conditional execution optimizations. Full article

Direction-of-arrival (DOA) estimation is fundamental in array signal processing, yet classical algorithms suffer from significant performance degradation under low signal-to-noise ratio (SNR) conditions and require computationally intensive eigenvalue decomposition. This study presents a systematic comparative analysis of two backbone networks, a convolutional neural network (CNN) and long short-term memory (LSTM) for DOA estimation, addressing two critical research gaps: the lack of a mechanistic understanding of architecture-dependent performance under varying conditions and insufficient validation using real measured data. Both networks are trained using cross-spectral density matrices (CSDMs) from simulated uniform linear array (ULA) signals. Under baseline conditions (1° classification interval), both CNN and LSTM methods reach an accuracy (ACC) above 98%, in which the error is ±1° for CNN and ±2° for LSTM, only existing in the end-fire direction. Key findings reveal that LSTM maintains above 90% accuracy down to −20 dB SNR, demonstrating superior noise robustness, whereas CNN exhibits better angular resolution. Four performance boundaries are identified: optimal performance is achieved at half-wavelength element spacing; SNR crossover occurs at −20 dB below which accuracy drops sharply; the snapshot threshold of 32 marks the transition from snapshot-deficient to snapshot-sufficient conditions; the array size of 8 is the turning point for the performance variation rate. Comparative analysis against traditional methods demonstrates that deep learning approaches achieve superior resolution ability, batch processing efficiency, and noise robustness. Critically, models trained exclusively on single-target simulated data successfully generalize to multi-target experimental data from the Shallow Water Array Performance (SWAP) program, recovering primary target trajectories without domain adaptation. These results provide concrete engineering guidelines for architecture selection and validate the sim-to-real generalization capability of CSDM-based deep learning approaches in underwater acoustic environments. Full article