Search Results (192)

General gaze following aims to infer the region attended to by a person within a natural scene and offers a computational perspective on visual attention and social scene understanding. Existing direction-guided methods usually reduce gaze direction to a single vector or spatial mask. This deterministic treatment can obscure directional ambiguity, discard coexisting candidate directions, and propagate early estimation errors to gaze target localization when head cues are weak or multiple targets are plausible. To address these limitations, we propose an uncertainty-aware framework based on Circular Direction Distribution Learning (CDDL) and Probabilistic Gaze Geometry Modeling (PGGM). CDDL represents gaze direction as a 72-bin circular probability distribution under von Mises soft supervision, thereby preserving neighboring directional hypotheses before target localization. Rather than predicting a target space distribution directly, PGGM aggregates the direction distribution into 18 groups and projects the retained hypotheses into cone-like spatial probability fields, allowing spatial tolerance to expand with distance while preserving a channel-wise direction-to-region representation. The full probability volume is fused with the RGB scene image at the input level and processed by a ResNet-50-FPN network for gaze heatmap prediction. Controlled experiments on GazeFollow demonstrate competitive localization and support the contribution of direction-distribution learning and channel-wise geometric projection. Zero-shot transfer, dataset-level uncertainty and failure analyses, efficiency measurements, and qualitative results further indicate interpretable behavior together with domain-dependent and computational trade-offs. Full article

Background/Objectives: Diffractive multifocal intraocular lenses (IOLs) are widely used following cataract surgery, but their impact on peripheral retinal image quality in ultra-widefield (UWF) fundus imaging has not been objectively characterized. This study aimed to quantify peripheral image sharpness in UWF fundus photographs and assess wavelength-specific differences between diffractive multifocal and enhanced monofocal IOLs. Methods: In this retrospective cohort study, 65 eyes from 51 patients implanted with enhanced monofocal (Tecnis Eyhance; n = 32) or diffractive multifocal IOLs (Tecnis Synergy, n = 19; Clareon PanOptix, n = 14) were analyzed. Non-mydriatic UWF fundus photographs (Optos; ~200°) were acquired within 1 month postoperatively. Red (R), green (G), and composite channel images were analyzed using three sharpness metrics—Laplacian variance, Brenner gradient, and Tenengrad gradient—within a predefined peripheral annular region (1.5–3.0 times the fovea–optic disc distance). Generalized estimating equations (GEE) with age as a covariate were applied as the primary analysis to account for inter-eye correlation and age differences between groups. Results: Brenner and Tenengrad gradient scores were significantly lower in the multifocal group in the R channel (p = 0.037 and p = 0.041, respectively), but not in the G or composite channels. The Laplacian metric revealed no significant intergroup differences in any channel. Subgroup analysis revealed no significant difference between hybrid multifocal and quadrifocal subtypes (p > 0.05). Group-level mean heatmaps confirmed a spatially distributed reduction in gradient magnitude across the annular zone in multifocal eyes. Conclusions: Diffractive multifocal IOL implantation was associated with a localized reduction in peripheral R channel sharpness, hypothesized to reflect transverse chromatic aberration. Composite channel quality remained comparable between groups, suggesting limited impact on routine clinical imaging. These findings may inform IOL selection and interpretation of peripheral retinal findings in multifocal IOL patients. Full article

To address three major industry pain points—the lack of quantified profit-sharing standards in the two-stage pricing under the separation model of military research and production, the absence of stable Nash equilibrium in single-layer synchronous optimization, and insufficient incentives for full-cycle process optimization in design units—this paper constructs a two-level Stackelberg leader–follower game model with the general contracting unit as the leader and design and general contracting units as followers. This aligns with current prototype incentives and phased pricing policies for production rewards and penalties. At the theoretical level, it improves the complete proof system for the two-stage concave profit two-level Stackelberg Nash equilibrium, distinguishes the mathematical differences in equilibrium existence between sequential decision-making and synchronous optimization, and extracts general rules for phased differentiated profit sharing: high-innovation segments should be allocated more profit weight; simply maximizing total alliance profit may cause imbalanced interests, while introducing a minimum net profit-weighted objective can achieve Pareto improvements without profit loss. This conclusion can be applied to multi-stage general contracting scenarios across industries, such as EPC and military–civil collaborative innovation, enriching the basic theory of profit sharing and hierarchical games. Theoretically, the existence of the lower-level Nash equilibrium is proven using Brouwer’s fixed-point theorem, and combining it with the strictly monotonically decreasing feature of the best response function, uniqueness of the equilibrium is derived. Multiple sets of differentiated initial values are simulated to rule out multi-equilibrium bifurcation risk. The model incorporates the military’s reward and penalty policies as rigid exogenous constraints, sets dual individual rationality constraints of ‘cooperative profit greater than baseline profit with no allocation, and both parties’ net profit non-negative,’ and introduces differentiated cost-reduction efficiency and quadratic increasing effort costs to characterize the heterogeneous input of the two types of development entities. For models with piecewise nonlinearity and multi-constraint nonconvex structures, this paper modifies the standard PSO into a Bi-PSO solving framework through hierarchical temporal adaptation. It does not innovate the underlying particle update mechanism and is only used to match the sequential decision order of the leader–follower game. By comparing five algorithms—IPM, GA, SA, DE, and adaptive PSO—through 20 repeated simulations: gradient-based interior point methods easily get stuck in locally invalid solutions that violate cooperation thresholds; differential evolution has the best numerical global search performance, but all general evolutionary algorithms optimize allocation and effort variables simultaneously, disrupting the Stackelberg hierarchical timing. Only Bi-PSO maintains consistent game logic. Using a pricing case for a certain type of equipment and jointly calibrating all parameters with policy documents, three simulation scenarios were set up: no allocation, equal 50/50 split, and single-layer profit maximization. Under the no-allocation mode, R&D investment from the design unit drops to zero and alliance benefits plummet; a blanket equal split ignores differences in technical contributions across two stages, leading to clear efficiency losses; single-layer optimization only pursues total profit maximization, causing a severe imbalance in profit distribution. The two-layer basic framework can achieve the upper limit of alliance benefits, and by adding a weighted optimization goal that considers both total profit and cooperation fairness, it can achieve equal net profits for both parties without reducing overall profit. Through single-parameter sweeps and two-factor heatmap simulations, the study further revealed the coupled effects of main party efficiency and mass production rewards and penalties on equilibrium input and optimal sharing ranges. A robust check was performed by replacing the logarithmic concave output function, producing a standardized allocation range resilient to parameter perturbations: optimal split for the prototype stage is 0.4–0.6 for the design unit, and for mass production stage 0.7–0.9. The findings suggest that high-contribution stages in multi-phase collaboration contracts should receive more benefits, and a weighted fairness objective can achieve Pareto improvements. These conclusions can extend to multi-stage collaboration scenarios such as EPC and military–civilian cooperation. Theoretically, this research further completes the equilibrium proof system for two-party concave payoff two-layer games, providing a new reference for the theory of phased differentiated benefit-sharing contracts in the military sector. Methodologically, it proposes a two-layer intelligent solving tool adapted to leader–follower sequential decisions, effectively mitigating issues where single-layer model equilibria fail or analytical algorithms struggle with multi-constraint nonconvex games. The results can provide quantitative support for the military, general contracting unit, and design unit in drafting equipment incentive pricing contracts and managing full-cycle cost collaboration. Full article

Gaze target detection requires understanding where a person is looking by jointly reasoning about the gazer and the surrounding scene. While recent methods have benefited from powerful pretrained visual backbones, they often treat gaze prediction as a generic localization problem and overlook a key property of the task: the target should be interpreted in relation to the person’s head. This limits their ability to model direction, distance, and head-scene dependencies in a unified manner. We propose GazeHRNet, a head-centric reasoning framework for RGB-based gaze target detection. Instead of relying on absolute image coordinates or auxiliary geometric inputs, GazeHRNet represents the scene from the gazer’s perspective through Head-Centric Polar Encoding and organizes visual features by their spatial relevance to the head via Head-Aware Attention Routing. It further combines coarse spatial reasoning with fine-grained anisotropic heatmap prediction, enabling reliable target localization under cluttered scenes and varying head positions. Experiments on GazeFollow and VideoAttentionTarget show that GazeHRNet achieves 0.952 and 0.929 AUC with L2 distances of 0.102 and 0.103, respectively, using only RGB input and 3 M trainable parameters. Cross-dataset evaluation further demonstrates improved robustness and generalization across different scenes and subject distributions. Full article

Introduction: Proper analysis of high-throughput lipidomic data requires specialized tools for data processing, normalization, visualization, and statistical and bioinformatic analysis. However, limitations in lipid parsing, data processing, and visualization capabilities in existing software packages create challenges for comprehensive lipidomic data analysis. To address these limitations, we developed LipidAnalyst (v 1.0.3), a user-friendly tool designed to facilitate efficient lipid parsing and processing, visualization, and analysis of lipidomic datasets. Methods: LipidAnalyst was developed using the R Shiny framework. It is hosted on MiServer for online work but can also be downloaded from GitHub. Results: LipidAnalyst provides functionalities in three major areas: data processing, visualization, and statistical analysis. Data processing features include quality control filtering, normalization, internal standard-based quantification, and unique capabilities for missing-value imputation, lipid parsing, and aggregation. Visualization tools include box and violin plots for data distribution assessment, principal component analysis (PCA) plots, hierarchical clustering and differential abundance heatmaps, volcano plots, correlation plots, and Debiased Sparse Partial Correlation (DSPC) clustering plots. Statistical analysis modules include t-test, analysis of variance (ANOVA), Partial Least Squares Differential Analysis (PLS-DA), Orthogonal Partial Least Squares Differential Analysis (OPLS-DA), and Random Forest (RF) modeling. Conclusions: LipidAnalyst is a comprehensive platform for optimal processing, visualization, and analysis of lipidomic data. By integrating advanced data processing workflows with extensive visualization and statistical analysis capabilities, LipidAnalyst enables researchers to explore lipidomic datasets more effectively and develop informed analytical strategies. Full article

Cementitious sand and gravel (CSG) is commonly used in construction engineering; however, its mix proportion design is complex, and traditional physical experiments face limitations such as long cycles, high costs, and susceptibility to external factors when obtaining high-quality sample data. In this study, a foundational dataset was first acquired through physical experiments: 100 sets of CSG specimens with different mix proportions (cement content 40, 50, 60, 70 kg/m³; water-to-binder ratio 1.0, 1.2, 1.4; sand ratio 0.1, 0.2, 0.3, 0.4; fly ash content 20, 30, 40, 50 kg/m³) were prepared. After 28 days of standard curing, compressive strength and splitting tensile strength tests were conducted using a WAW-1000 electro-hydraulic servo universal testing machine, yielding 100 sets of real mechanical property data. The coefficients of variation for all test groups were below 10%, confirming the reliability and repeatability of the experimental data. On this basis, a data augmentation method based on Conditional Tabular Generative Adversarial Networks (CTGAN) is proposed. Through adversarial training between the generator and the discriminator, the model learns the multi-dimensional distribution characteristics of the original CSG data and generates 100 synthetic samples, which are then merged with the original data to expand the dataset to 200 samples. The quality of the synthetic data is evaluated using Wasserstein distance and correlation matrix heatmaps. Furthermore, a hybrid XGBoost–LSTM prediction model is proposed—XGBoost is used for feature construction to capture nonlinear interactions among mix proportion variables, and the constructed features are then fed into an LSTM network for sequential learning and regression prediction. The results show that the CTGAN-generated data are highly consistent with the original data in terms of kernel density distributions and variable correlations, with Wasserstein distance significantly superior to four comparative methods: Bootstrap, SMOTE, GaussianCopula, and TVAE. After augmentation, the XGBoost–LSTM model achieves a coefficient of determination (R²) of 0.9897 for compressive strength prediction (vs. 0.9793 before augmentation) and 0.9801 for splitting tensile strength (vs. 0.9882 before augmentation, a slight decrease). The mean absolute percentage errors (MAPE) are 4.49% and 4.11%, and the root mean square errors (RMSE) are 0.201 and 0.049, respectively; both error metrics are reduced compared with those before augmentation. Compared with baseline models including XGBoost, LSTM, Random Forest (RF), and Support Vector Regression (SVR), the XGBoost–LSTM model exhibits the best performance across all evaluation metrics, and Wilcoxon signed-rank tests confirm that the performance differences are statistically significant (p < 0.05). The proposed method of CTGAN-based data augmentation combined with the XGBoost-LSTM hybrid model provides an effective solution to the problem of insufficient CSG sample data and offers a reference for data enhancement and performance prediction of other small-sample materials. Full article

This study presents two workflows for leveraging longitudinal computed tomography (CT) datasets when developing deep learning-based detection systems for gradually developing postharvest disorders. Workflow 1 (Longitudinal Benchmarking) benchmarks neural networks by training and testing them on images from different stages of disorder progression. It examines the trade-off between detecting a disorder early or accurately and evaluates whether neural networks can generalize across time points. Workflow 2 (Longitudinal eXplainable Artificial Intelligence (XAI) Heatmaps) provides heatmaps that indicate how changes over time affect the outcomes of neural networks. It uses image registration to align an earlier-acquired image and then uses it as a baseline when calculating the heatmap. The workflows are demonstrated on a dataset of ‘Braeburn’ apples that were CT-scanned multiple times while developing internal browning during controlled-atmosphere (CA) storage and shelf life. The Longitudinal Benchmarking workflow was used to investigate whether images acquired immediately after CA storage can be used to predict the eventual browning after a shelf-life period, which is highly relevant in industrial practice. Moreover, the longitudinal XAI heatmaps avoided artifacts caused by out-of-distribution baselines or identical baseline regions, which occurred with conventional black or zero baselines. Full article

In millimeter-wave frequency-modulated continuous-wave (FMCW) multiple-input multiple-output synthetic-aperture radar (MIMO-SAR) imaging, platform displacement beyond the spatial Nyquist limit during a slow-time sampling interval creates aperture gaps, causing azimuth aliasing and degraded resolution. This paper proposes a spectral-prior-guided Swin TransUnet (SSTU) method for suppressing azimuth ambiguity in sparse moving-array imaging. Gaussian soft labels derived from point-scatterer positions formulate localization as heatmap regression and guide mainlobe learning. A two-dimensional fast Fourier transform (2D-FFT) layer then constructs a range–azimuth spectrum that exposes main peaks, sidelobes, and periodic grating lobes. A convolutional encoder extracts local spectral features, Swin Transformer blocks model long-range ambiguity correlations, and a U-Net-style multiscale decoder reconstructs high-resolution range–azimuth images. Simulations show that SSTU reliably recovers multiple point targets from noise and grating lobes despite substantial aperture gaps. At 60% aperture sparsity and signal-to-noise ratio (SNR) above −6 dB, it achieves a root mean square error (RMSE) below ${10}^{-2}$ and an azimuth ambiguity suppression ratio better than −30 dB, outperforming conventional methods. Measurements using a 77 GHz radar platform further demonstrate high-quality outdoor imaging of randomly distributed strong scatterers at 60% moving-aperture sparsity. Full article

Phytobiotics are promising dietary tools to improve metabolic stability and physiological resilience against chemical stressors in aquaculture. This study evaluated nrf2-linked antioxidant and metabolic modulation by dietary Origanum vulgare essential oil (OEO) in Nile tilapia (Oreochromis niloticus) under malathion-induced organophosphate stress. Fish, with an initial weight of 5.76 ± 0.56 g, were distributed in three tanks per treatment, with 15 fish per tank, and fed diets containing 0 (control), 0.75, 1.50, 2.25 and 3.00 g kg⁻¹ OEO for eight weeks, followed by exposure to malathion (7.04 mg L⁻¹) for 96 h. OEO enhanced (p < 0.05) growth performance, feed utilisation, and survival after exposure, with the greatest productive and survival outcomes at 3.00 g kg⁻¹. Fish that received the supplement also had lower (p < 0.05) lipid peroxidation, better antioxidant enzyme activity and a more favourable modulation of nrf2, gpx and keap1 expression, together with biochemical and histological patterns that were consistent with a better condition of the liver and kidney. The heatmap and PCA supported a treatment-related separation, with 2.25 and 3.00 g kg⁻¹ OEO showing the most favourable integrated physiological profiles. Dietary OEO was associated with antioxidant, metabolic and tissue-level resilience linked to modulation of nrf2-related transcriptional responses. Full article

In the extremely arid Gobi environment, it remains unclear whether afforestation with saxaul (Haloxylon ammodendron) is an effective ecological restoration strategy or whether it may trigger ecological risks under severe water limitation. This study examined saxaul shelterbelts of different ages across precipitation gradients (0–50, 50–100, and 100–150 mm·yr⁻¹) in the Alxa Gobi based on 48 plots. It systematically assessed the effects of precipitation on stand growth, soil particle-size distribution, and nutrient dynamics inside and outside plantations, and used partial least squares path modeling (PLS-PM) to analyze the coupling mechanisms among precipitation, soil, and plant growth. The results showed that precipitation was the key factor controlling the survival and growth of saxaul, while planting density further regulated survival rates within different precipitation zones. Plant height, crown width, and basal diameter generally showed better performance in higher-precipitation zones, although differences among plantation ages may have been influenced by variation in initial planting density. Among the three precipitation zones, plants in the 100–150 mm zone exhibited the best growth performance. In low-precipitation areas (≤50 mm), the growth of saxaul was strongly limited, and afforestation disturbance was associated with disruption of the surface gravel layer, soil coarsening, and inadequate nutrient accumulation. In contrast, in medium- and high-precipitation areas (50–150 mm), saxaul plantations established at appropriate densities are more conducive to the accumulation of fine soil particles and nutrient enrichment. Correlation heatmaps and PLS-PM results further showed that precipitation gradient, soil texture, soil fertility, and plant growth were closely coupled. Moreover, the associations between soil and vegetation variables were stronger inside the shelterbelts than outside the shelterbelts, indicating that more pronounced local soil–vegetation feedbacks may have been formed after the establishment of artificial Haloxylon ammodendron stands. Overall, the suitability of saxaul plantations in the Alxa Gobi showed clear precipitation-dependent differentiation, with approximately 50 mm representing a practical lower limit for saxaul plantation establishment. Large-scale saxaul plantations is not recommended in areas with precipitation ≤ 50 mm, where low-disturbance restoration focused on gravel-layer protection should be prioritized; in contrast, areas receiving 50–150 mm precipitation are more suitable for plantation establishment under appropriate density control. These findings provide a scientific basis for sustainable afforestation, regional allocation, and low-disturbance management in extremely arid Gobi regions under the principle of matching vegetation restoration to water availability and site conditions. Full article

Lactiplantibacillus plantarum is a highly versatile lactic acid bacterium, widely distributed across diverse ecological niches. Although often described as a nomadic species, increasing evidence suggests that strains from different habitats may retain niche-associated functional traits. This study investigated the phenotypic biodiversity of forty L. plantarum strains isolated from four ecologically distinct environments: wine, honeybee gut, trout intestine, and pre-weaning infant feces. Growth performance at different temperatures and on various carbon sources, acidification capacity, and β-glucosidase activity were evaluated and integrated using multivariate statistical analyses. Significant differences in β-glucosidase activity were observed among ecological groups (Kruskal–Wallis, p = 0.001), with wine-associated strains exhibiting the highest enzymatic activities and trout-derived isolates the lowest. Growth and acidification traits showed more limited variation among habitats, indicating that these physiological characteristics are largely conserved within the species. Heatmap visualization, principal component analysis (PCA), and hierarchical clustering revealed substantial phenotypic heterogeneity among strains. PCA indicated that growth performance and acidification traits contributed primarily to the first principal component, whereas β-glucosidase activity and differential fructose utilization were major contributors to the second component. Permutational multivariate analysis of variance (PERMANOVA) confirmed a significant effect of ecological origin on the overall phenotypic structure (p = 0.006), although habitat explained only 15.3% of the total variance (R² = 0.153). Overall, the results show that ecological origin contributes to the phenotypic diversification of L. plantarum populations while preserving the extensive functional versatility characteristic of this species. β-Glucosidase activity emerged as the most discriminating phenotypic trait among ecological groups and represented the principal niche-associated functional signature identified in this study. Full article

Abandoned pastoral settlements can create disturbed and nutrient-enriched microsites that favor the dominance of native expansive plant species. Yet limited empirical evidence exists on how abandoned Maasai bomas influence the spread of Urtica massaica and associated plant community changes in the Ngorongoro Conservation Area, a UNESCO World Heritage Site in Tanzania. This study assessed the influence of abandoned bomas on plant species abundance, richness, diversity, soil seedbank status, and the spatial distribution of U. massaica. A multistage stratified random sampling design was used, whereby Ngorongoro and Nainokanoka zones were selected from the designated management zones of the conservation area. Vegetation data, soil samples, and geographic coordinates of abandoned bomas were collected from abandoned boma sites and adjacent control sites. Plant species abundance, richness, and diversity were compared between abandoned bomas and control sites after testing data normality using the Shapiro–Wilk test. Independent-sample t-tests were used for normally distributed data, while Wilcoxon rank-sum tests were used for non-normally distributed data. Spatial distribution of U. massaica was assessed using GIS-based heatmap and kernel density estimation. Results showed that native plant species abundance was significantly higher in control sites than in abandoned bomas. Plant species richness and diversity also differed significantly between sites in both Ngorongoro and Nainokanoka, with control sites supporting higher richness and diversity. However, soil seedbank results showed no significant differences in species richness and diversity between soils collected from abandoned bomas and control sites, although slightly higher values were observed in control soils. Spatial analysis revealed that U. massaica hotspots were concentrated mainly in highland areas with high densities of abandoned bomas. These findings suggest that abandoned bomas may act as focal points for U. massaica establishment and dominance, reducing aboveground plant diversity while retaining some potential for natural regeneration through the soil seedbank. Management interventions should prioritize abandoned bomas as key sites for controlling U. massaica spread and supporting vegetation recovery in the Ngorongoro Conservation Area. Full article

Stilbocrea banihashemiana Bolboli, Tavakolian & Mostowf. is an emerging pathogen causing canker and dieback in a broad range of fruit and ornamental trees in Iran, and its distribution is expanding across the country. Extensive surveys conducted over five consecutive years (2019–2023) yielded 88 isolates of S. banihashemiana from multiple hosts, including different fig (Ficus caricae L.) cultivars, as well as loquat (Eryobotria japonica (Thunb.) Lindl.), pomegranate (Punica granatum L.), and walnut (Juglans regia L.) trees, across eight distinct regions of southern Iran. Species identification was performed morphologically and molecularly by employing the S. banihashemiana-specific primer pair TEF-Sb1 and TEF-Sb3. The genetic diversity of the S. banihashemiana population of isolates was assessed using eight inter-simple sequence repeats (ISSRs) markers. The UPGMA dendrogram demonstrated broad genetic variability among the isolates, with similarity coefficient values spanning from 0.46 to 1.00. This wide range indicates the presence of multiple divergent genotypes within the population, rather than a single dominant lineage. Principal coordinate analysis (PCoA) grouped the 88 isolates into three distinct genetic clusters that partially corresponded to geographic origin and host species. Pathogenicity assessment of 53 selected isolates from various hosts and geographic origins on detached fig shoots demonstrated highly significant variability in aggressiveness among isolates originating from different host species and geographically distinct regions. Multivariate analysis using principal component analysis (PCA) combined with heatmap-based clustering of the aggressiveness dataset clearly separated the isolates into four distinct groups, ranging from highly to less aggressive. A susceptibility assessment of 10 fig cultivars using the ex-type-isolate of S. banihashemiana revealed that the pathogen caused internal lesions and wood discoloration in all cultivars. Based on statistical analysis, the cultivars were classified into three groups: susceptible (cv. ‘Siah’), moderately susceptible (‘Brown Turkey’, ‘C8-M’, ‘C8-F’, ‘Dehdez’, ‘Gilasi’, ‘Payves’, ‘Shah-Anjeer’ and ‘Sabz’), and less susceptible (‘Matti’). High genetic variability, multiple-host association, and partial geographic structure indicate that in Fars Province S. banihashemiana’s population structure and epidemiology are complex, with high adaptive potential. This complexity may influence disease spread, management strategies, and long-term evolutionary trajectories. Full article

Driving behavior recognition plays a crucial role in intelligent driving systems and road traffic safety. Due to insufficient driving experience and limited ability to allocate visual attention, novice drivers are considered a high-risk group for traffic accidents. Existing approaches primarily focus on experienced drivers and rely on single-modal eye-tracking data, making it difficult to model spatial attention distributions and long-term temporal dependencies simultaneously. Moreover, these methods are often affected by modality asynchrony during multimodal fusion, further limiting performance gains. To address these challenges, this study proposes a novice driver behavior recognition method based on bimodal eye-tracking features and a gated cross-modal attention fusion (GCMAF) mechanism. The model adopts a spatial–temporal dual-branch architecture. The spatial branch employs ResNet34 to extract eye-tracking heatmap features to represent the visual attention distribution. In contrast, the temporal branch integrates a 1D-CNN with the Mamba model to capture local dynamic patterns and long-range temporal dependencies. In the fusion stage, the GCMAF module is introduced to enhance cross-modal interactions, and a gating mechanism is further used to adaptively adjust modality weights, thereby mitigating the adverse effects of modality asynchrony. To validate the effectiveness and generalization ability of the proposed method, repeated experiments and five-fold cross-validation are conducted. The results demonstrate that the model achieves an average classification accuracy of 93.86% across four driving behavior categories, with standard deviations below 0.3%. Compared with baseline methods, paired t-test results show that the performance improvement is statistically significant (p < 0.01). Ablation studies further confirm the independent contribution of each component. Overall, the proposed method outperforms existing approaches in terms of accuracy and stability, providing effective support for driving behavior assessment and proactive safety warning systems. Full article

Landscape Visual Assessment (LVA) has long examined how landscape elements influence visual perception and aesthetic response, yet the question of which elements attract attention remains underexplored in historic parks. Compared with urban parks, streetscapes, or natural landscapes, historic parks are shaped by a distinctive combination of natural features, cultural structures, and historically embedded symbolic meanings. In this study, we categorize landscape elements in historic parks into Natural Landscape Elements (NLE) and Cultural Landscape Elements (CLE), and investigate their relative visual salience in three historic parks in Beijing: Taoranting Park, the Summer Palace, and Beihai Park. We adopt a multi-method exploratory framework integrating cognitive maps, eye-tracking experiments, computational vision analysis, and GIS/UGC-based spatial analysis. The study draws on 30 cognitive maps, participant-level eye-tracking data from 30 valid participants, supplementary heatmap-based computational image analysis, and large-scale geotagged photo-density data. The results show that, within the exploratory sample, CLE were more frequently associated with strong memory impressions than NLE. In the pooled eye-tracking analysis, CLE showed higher attention scores overall, but this pattern was not stable across all parks and was strongly context-dependent. The computational and spatial analyses further suggest that attention distribution is influenced not only by the presence of specific elements, but also by color contrast, the spatial coupling of CLE and NLE, and the broader organization of park scenes and visitor activity zones. Rather than proposing a universal model of visual attention in historic parks, this study offers an exploratory, context-sensitive account of how cultural and natural landscape elements jointly shape attention across perceptual and spatial scales. The findings contribute to landscape visual assessment research by extending it into the specific setting of historic parks and by demonstrating the value of combining perceptual, computational, and spatial methods in a complementary framework. Full article