Search Results (798)

Diffusion-based generative models achieve high-fidelity image synthesis; however, controlling internal representations for abstract visual concepts remains challenging due to the ambiguity of textual descriptions. In this work, we propose a prompt-guided concept-vector learning framework for the controllable manipulation of such concepts without requiring external human-annotated image pairs, segmentation masks, identity labels, or manually annotated editing targets. The method introduces a learnable concept vector optimized in the bottleneck (mid-block) feature space of a pretrained Stable Diffusion U-Net, while keeping all pretrained model parameters frozen. A multi-prompt data generation strategy based on paired positive and neutral prompts provides weak semantic guidance for capturing the target concept direction and reducing dependence on a single prompt formulation. The learned vector is further applied in an image-to-image setting through controlled noise injection and concept-guided denoising, enabling the semantic modification of real images while preserving structural content. The concept strength is controlled by a scaling parameter $\gamma$, while the image-to-image noise strength is controlled by $\beta$, allowing for a practical balance between semantic modification and structural fidelity. Experiments are conducted on two main abstract concepts, perfect skin and peaceful lake, with additional qualitative analysis on subjective portrait-level concepts. Quantitative evaluation using SSIM, LPIPS, and CLIP similarity demonstrates that the proposed method improves semantic alignment while maintaining structural preservation compared with Stable Diffusion image-to-image baselines. A human preference study further shows that concept-injected outputs are preferred in 76.0% of responses for perfect skin and 85.7% for peaceful lake. Ablation studies further demonstrate the controllability and robustness of the proposed framework. Overall, the method provides a simple and parameter-efficient approach for interpretable concept-level manipulation in diffusion models. Full article

Background: Orthodontically induced external root resorption (ERR) may affect long-term tooth stability, requiring a reliable assessment of root length changes. This study developed an artificial intelligence (AI)-based framework for automatic segmentation and quantitative evaluation of root resorption, comparing fixed appliances and clear aligner therapy. Methods: A dataset of 100 anonymised orthopantomographic (OPG) radiographs (50 fixed appliance, 50 clear aligner) obtained before and after orthodontic treatment was analysed. A U-Net model was trained for automatic tooth segmentation and quantitative assessment of tooth length changes. Measurements were computed from both AI-predicted and Dentist-annotated masks using pixel-based Python analysis (Python version 3.6), and pre-post differences were compared between methods. Results: The segmentation model achieved high performance with Intersection over Union values up to 89% and Dice Similarity Coefficients up to 95%. Quantitative analysis demonstrated significant reductions in root length following orthodontic treatment in both modalities (p < 0.0001). In the fixed appliance group, AI-based measurements showed an average root length reduction of 8.95%, while human measurements indicated a reduction of 10.16%. In the clear aligner group, AI measurements demonstrated a reduction of 2.81%, compared with 4.98% in human measurements. Root resorption was significantly greater in the fixed appliance group than in the clear aligner group (p < 0.0001). AI-derived measurements showed strong agreement with expert evaluations. Conclusions: AI-based analysis of panoramic radiographs may provide a reliable and reproducible approach for quantifying orthodontically induced root resorption. The findings suggest that clear aligner therapy was associated with lower root length reduction; however, larger multi-centre studies and external validation are required before clinical implementation. Full article

Immunohistochemistry (IHC) is a widely used method for detecting specific proteins in tissue samples, helping diagnose diseases such as cancer. Traditional analysis methods rely heavily on human interpretation, which can lead to inconsistencies. In this study, we propose DeepFlare, a weakly supervised deep learning framework for cross-modality translation and segmentation of immunofluorescence and immunohistochemistry images. The proposed method utilizes multiplex immunofluorescence (mpIF) and co-registered IHC images, combined with preprocessing techniques such as affine transformation, stain normalization, noise reduction, and artifact removal. Multiple imaging channels, including hematoxylin, DAPI, Lap2, and nuclear envelope signals, are leveraged to generate segmentation masks using a U-Net++ architecture. The final segmentation mask is obtained through weighted fusion of modality-specific outputs. A generative adversarial network (GAN) is employed to measure translation fidelity between generated and real images. Weakly supervised learning techniques, including image-level supervision and consistency constraints, are applied to enhance performance under limited annotation scenarios. Pretrained pathology foundation encoders such as UNI and Virchow are integrated to extract multi-scale morphological and contextual features. Explainable AI techniques are incorporated to highlight critical regions and refine model attention. Experimental results demonstrate strong performance, achieving an SSIM of 0.7077 for image translation and a Dice score of 0.7424 for segmentation. The integration of the UNI encoder provides marginal improvement over the baseline (0.72 Dice score), indicating limited domain adaptation without fine-tuning on the dataset of 1264 training samples. Full article

The rapid progress of generative speech synthesis and voice-cloning technologies has enabled the creation of highly natural synthetic voices that pose a serious threat to telecommunication security. While most prior studies evaluate human ability to detect audio deepfakes using high-quality, studio-grade recordings, little is known about how real-world telecommunication channels affect perceptual detection. This study investigates the influence of three transmission scenarios—GSM (AMR-NB), VoLTE (AMR-WB), and VoIP with packet-loss modeling—on the human ability to distinguish natural speech from AI-generated speech. A custom speech corpus was developed, consisting of natural recordings from nine speakers and corresponding synthetic utterances generated using a state-of-the-art voice cloning system (ElevenLabs). All samples were processed through simulated telecommunication channels using real codec implementations. A listening test with 95 participants was conducted, involving binary classification (human vs. synthetic) and confidence ratings. Results show an overall detection accuracy of 54.8%, confirming that humans are poorly equipped to identify synthetic speech. Surprisingly, the highest accuracy was achieved for the narrowband GSM channel (63.7%), while VoLTE yielded the lowest performance (44.0%). The findings suggest that restricted bandwidth may emphasize prosodic irregularities typical of generative models, whereas high-quality channels mask synthetic artifacts, increasing susceptibility to voice spoofing. The results highlight the necessity of deploying additional security mechanisms in telecommunication systems relying on voice identity verification. Full article

This article examines how Ecuadorian national digital newspapers represented adolescents and youth-coded young adults associated with crime during 2025, with particular attention to lexical labeling, moral attribution, visual framing, editorial prominence, news values and the near-presence or absence of educational discourse. The study is based on qualitative content analysis of Spanish-language digital press coverage published in El Universo, El Comercio, Extra, La Hora, GK, Primicias, Vistazo, El Mercurio and Expreso across seven journalistic genres: news, note, feature article, report, editorial, interview and chronicle. The article argues that media discourse does not merely describe youth violence; it actively constructs public intelligibility about who young people are, how danger is recognized and whether social responses are imagined in punitive, preventive or restorative terms. Grounded in media framing theory, news values, moral panic studies, child-friendly justice, critical sociology, school push-out scholarship and philosophies of education and human development, the article shows the inferential route from media representation to educational reintegration: when coverage individualizes adolescent violence, minimizes school interruption and masks structural conditions, it narrows the policy imagination through which young people are understood as educable, rights-bearing and recoverable subjects. The paper ultimately argues that the long-term reduction of violence in Ecuador requires not only security responses but also an integral reintegration agenda centered on education, dignified work, child-sensitive justice and restorative social policy. Full article

Modern fulfillment centers share floor space with human workers, making warehouse multi-robot task allocation a safety-critical problem. We propose SHARP (Safe Heterogeneous Allocation with Risk Prediction), a Transformer-based constrained reinforcement-learning framework with a closed-form deployment-time safety mask. Under a Gaussian pedestrian belief and fixed closest-approach directions, the mask uses Bonferroni-allocated per-pair CVaR scores; a nonnegative mask score implies a conservative trajectory-level chance constraint under the stated assumptions. We also present an idealized primal–dual surrogate analysis, without claiming global convergence for the nonconvex Transformer/PPO implementation. Expanded experiments use ten training seeds per learned method and deterministic final-checkpoint evaluation on twenty independently generated held-out instances. No statistically significant difference between SHARP and Lagrangian-PPO was detected in any of the four scenarios. The held-out analysis further reveals late-training instability and severe over-conservatism in the dense S40_high scenario. These findings position SHARP as an auditable geometric filtering mechanism, while identifying conservatism and training stability as important limitations for deployment. Full article

This research proposes a conditional diffusion-based workflow for early-stage floor plan design in university research buildings, addressing complex functional organization, strict boundary constraints, and quantitative area control. The method performs denoising directly in two-dimensional grid space and coordinates building outlines and functional area proportions through dual-condition injection using boundary masks and functional area matrices. A two-stage generation mechanism first constructs horizontal circulation and then generates the complete layout, while a statistic-network-guided explicit constraint improves global area consistency. Based on 600 standard-floor samples and an independent test set of 10 real projects, the method is evaluated through model comparison, ablation, and double-blind experiments. The results show that the proposed model achieves the best overall performance, with an FID of 50.3, a building boundary IoU of 99.9%, and horizontal circulation connectivity of 89.8%. The ablation results confirm that the two-stage mechanism and explicit statistical constraint substantially improve generation success and reduce area error. The expert evaluation indicates that AI-generated floor plans approach real cases in functional spatial form and design inspiration, although spatial organization rationality still requires improvement. The generated layouts can be converted into layered DXF files, supporting subsequent editing and human–AI collaborative design. Full article

Fusarium mycotoxins pose a major challenge for agriculture and the food industry due to their frequent occurrence in cereals. In addition to conventional mycotoxins, modified mycotoxins, including the subgroup of masked mycotoxins, are receiving increasing attention. These compounds are formed through plant defence mechanisms, food processing or biological transformations and are often undetectable using conventional analytical methods. Due to their potential reactivation in the digestive system of humans and animals, masked mycotoxins represent a hidden threat to food safety. This article examines the mechanisms of formation of modified mycotoxins, their occurrence in the food chain and their potential health risks. Particular emphasis is placed on the analytical methods required for their detection, including advanced chromatographic and spectrometric techniques. Understanding modified mycotoxins is crucial for the development of more effective control and prevention strategies. Improved agronomic practices, proper storage and advances in detection methods are essential to reduce exposure to these compounds and ensure food safety. This study provides a comprehensive overview of the current state of research on modified mycotoxins and underlines the need for further scientific research and regulatory guidance to protect consumer health and maintain confidence in the food industry. Full article

Traditional high-precision canopy segmentation heavily relies on tedious pixel-level manual annotation, while general-purpose zero-shot visual detection algorithms are prone to boundary adhesion and excessive computational load in dense forest areas. To address this, this study proposes a human–machine collaborative, efficient canopy segmentation and canopy cover inversion paradigm, combining the zero-shot pre-annotation capabilities of text-driven object detection with the high-precision segmentation advantages of the lightweight proprietary network LGBU-Net. In the offline annotation stage, this method automatically locates candidate canopy regions using Grounding DINO combined with text prompts and generates initial pixel-level masks using SAM. A high-quality training set is then constructed through minimal manual correction, significantly reducing the cost of traditional fully manual annotation. Subsequently, an improved LGBU-Net designed for complex forest conditions is used for supervised learning. In the feature extraction stage, a lightweight phantom-coordinate attention module (LG-CAM) is introduced to enhance the network’s focus on the geometric center of the tree canopy and suppress semantic interference caused by the forest background, light spots, and shadows. In the decoding stage, a boundary difference fusion module (BDF-Block) is deployed to alleviate the problem of adjacent tree canopy boundaries adhering by utilizing high-frequency gradient information from the underlying layers of UAV imagery. Combined with a boundary-aware hybrid loss function, the clarity of individual tree boundaries is further improved in the gradient domain. Experiments based on UAV imagery of high-density mixed and coniferous forests in Baishan, Jilin Province, show that, with low manual annotation costs, LGBU-Net achieves a canopy segmentation IoU of 90.45% and an individual tree separation F1 score of 89.35%, significantly outperforming general visual algorithms with zero-shot direct inference, and with only 4.85 M model parameters. Furthermore, the segmentation results are used for plot-level canopy vertical cover (CC) inversion, and the estimated values are highly consistent with ground-based measurements. This research provides a high-precision, low-annotation-cost technical solution with good edge deployment potential for large-scale forest resource surveys and forest understory light environment assessment. Full article

Most plant-based essential oil repellent products currently available on the market utilize a “green” approach based on the volatile properties of essential oils. In general, these essential oils contain terpenes, terpenoids, phenylpropanoids or benzenoids that can be used to either (1) eliminate a human’s scent through a process called odor masking, or (2) interfere with an insect’s ability to detect a person’s scent through interaction with both olfactory receptors and odorant binding proteins. Additionally, many of the essential oil blends that have been developed have been shown to exhibit antimicrobial and therapeutic properties. The primary drawback to using essential oil-based repellents is that their protection times vary widely, and typically last only a short period of time due to the volatile nature of the active ingredients, as well as differences in concentration and formulation among products. Encapsulation, nano-delivery systems, and rationally designed blend combinations are being proposed as potential methods to delay the release of the essential oil active ingredients, thus extending the duration of effectiveness of the repellent product. Since essential oils represent complex mixtures, there is a possibility that resistance to the repellent active ingredients could develop differently than it would for single-active agents. However, before such resistance can be assessed, the repellents must undergo extensive safety evaluations, along with standardized efficacy assessments against Environmental Protection Agency (EPA)-approved repellent products, and ultimately, field trials must be conducted in areas where the repellents will be used to prevent vector-borne diseases. In addition to conducting these evaluations, the repellents must comply with existing state and federal pesticide regulations. Full article

Intelligent perception of watermelon targets in complex field environments is a key prerequisite for automated harvesting and future collaborative decision-making with agricultural large models. To address severe leaf occlusion, large pose variation, dense adhesion among adjacent fruits, and the inability of conventional horizontal bounding boxes to accurately represent target orientation under natural cultivation conditions, this paper proposes an improved YOLO11n-OBB-based method for rotated watermelon detection. During data preparation, a semi-automatic annotation strategy combining segmentation-mask assistance with circumscribed rectangle fitting was adopted to efficiently construct a watermelon OBB dataset that closely matches the true physical boundaries of the fruits. On this basis, three structural improvements were introduced to the YOLO11n-OBB baseline: an LSK module was selectively embedded into the middle and later stages of the backbone to enhance adaptive receptive-field modeling and occlusion reasoning in complex bac kgrounds; the original neck structure was replaced with a lightweight BiFPN to strengthen bidirectional feature fusion for targets with large-scale variation in field scenes; and KFIoU Loss was incorporated into the rotated box regression branch to alleviate angle sensitivity and boundary discontinuity, thereby improving the convergence stability of orientation parameter learning. On the constructed watermelon OBB test set, the improved model raised mAP@0.5 (OBB) from 0.871 to 0.931, mAP@0.5:0.95 (OBB) from 0.670 to 0.736, Precision from 0.885 to 0.931, and Recall from 0.849 to 0.908 relative to the YOLO11n-OBB baseline (relative gains of 6.89%, 9.85%, 5.20%, and 6.95%, respectively), while keeping the inference speed at 100 FPS and the parameter count at only 2.71 M. While maintaining a compact model size and high real-time performance, the proposed method significantly improved rotated detection accuracy in crowded and overlapping scenes. In addition, the detection results were encapsulated into a structured JSON perception interface, preliminarily demonstrating the integration pathway of this lightweight front-end for task planning and human–machine collaborative operations with agricultural large models, and indicating its potential for future intelligent agricultural decision-making. Full article

Objectives: This study aimed to evaluate the influence of panoramic radiograph quality on the performance of an artificial intelligence (AI)-based tooth detection system and to identify specific image quality criteria associated with detection accuracy. Methods: A total of 424 panoramic radiographs were retrospectively selected from a clinical database. Radiographic quality was assessed using a modified Clinical Image Evaluation Chart, including criteria related to bite block presence, anteroposterior positioning, occlusal plane curvature, patient movement, anatomical coverage, overlapping contact points, air gap, contrast, cervical spine overlap, symmetry of the ascending mandibular ramus, and the number of visible teeth. Automated tooth detection was performed using a convolutional neural network based on the Mask R-CNN architecture (SynbrAIn, Italy). AI detection outputs were validated against expert human evaluation. Spearman’s rank correlation analyses were conducted to assess associations between individual image quality criteria and the number of AI detection errors per radiograph. Results: Significant negative associations were observed between AI detection errors and the number of visible teeth (ρ = −0.311, p < 0.001), presence of a bite block (ρ = −0.248, p < 0.001), reduced patient movement (ρ = −0.204, p < 0.001), correct anteroposterior positioning (ρ = −0.165, p < 0.001), and overall image quality score (ρ = −0.120, p = 0.010). In contrast, the presence of an air gap above the anterior teeth (ρ = 0.099, p = 0.042) and overlapping contact points (ρ = 0.122, p = 0.012) were positively associated with increased detection errors. No significant associations were identified for occlusal plane curvature, contrast, cervical spine overlap, anatomical coverage, or mandibular ramus symmetry. Overall, the AI system was more sensitive to indicators of anatomical completeness and patient positioning than to minor radiographic imperfections. Conclusions: Panoramic radiograph quality, particularly indicators of anatomical completeness and patient positioning, is associated with the performance of AI-based tooth detection. While the AI system demonstrated robustness to common image quality variations, adherence to standardized acquisition protocols remains important to minimize detection errors. Full article

Crashes involving freight and commercial vehicles impose substantial human and economic costs, yet most severity studies pool vehicle types or focus exclusively on heavy trucks, masking class-specific risk mechanisms. This study estimates separate Random Parameters Binary Logit models with heterogeneity in means and variances for three vehicle categories—heavy-duty multi-axle trucks (n = 6512), two-axle trucks (n = 2656), and light-duty pickup trucks (n = 23,477)—using 32,645 crash records from Thailand’s national highway network (May 2022–December 2024). Pairwise transferability tests rejected parameter transferability, with four of six comparisons exceeding the 97 percent confidence level (three of these above 99 percent; χ² = 85.38 to 240.01), confirming that disaggregate estimation is statistically warranted. Three core findings emerge: First, although barrier medians, cut-in-front maneuvers, and sideswipe crashes affect severity in consistent directions across all vehicle types, their magnitudes differ sharply: the protective effect of barrier medians is nearly six times larger for two-axle trucks (ME = −0.160) compared to heavy-duty trucks (ME = −0.028). Second, several determinants are class-specific: dark unlit conditions elevate severity only for two-axle trucks (ME = 0.128), flush medians only for heavy-duty trucks (ME = 0.040), and raised medians only for light-duty pickups (ME = 0.042). Third, no random parameter is common to all three models. Pooled models, therefore, impose misleading homogeneity assumptions; vehicle-type-specific estimation is essential for targeted safety policy. Full article

In arid inland watersheds, the compounding impacts of climate change and intensive human activities have severely altered hydrological regimes and accelerated landscape degradation. However, conventional spatial planning often overlooks the critical coupling between subsurface hydrological processes and surface landscape dynamics. Taking the Manas River Watershed in northwestern China as a representative case, this research investigates the multi-scale dynamics of landscape patterns and their underlying spatial determinants. Integrating multi-period land-use data (2000–2020), landscape metrics, and the GeoDetector model, we diverge from conventional uniform buffer approaches by redefining riparian boundaries utilizing four distinct River–Groundwater Transformation (RGT) patterns. This methodological shift reveals critical eco-hydrological heterogeneities previously masked by fixed-width approaches. Our multi-scale analyses demonstrate that watershed-level landscapes exhibited a trajectory of declining diversity, transient recovery, and ultimately, intensified fragmentation, while riparian patches concurrently expanded and became increasingly homogenized. GeoDetector assessments indicate a fundamental shift in driving forces: early-stage variations were constrained by natural factors, whereas post-2010 dynamics became overwhelmingly dominated by socio-economic determinants, particularly agricultural expansion and GDP growth. Crucially, our RGT-coupled spatial analysis reveals a strong spatial association between agricultural sprawl and landscape risk hotspots concentrated within groundwater overflow zones—a pattern consistent with, but not directly demonstrating, disrupted vertical hydrological connectivity. Direct verification of subsurface mechanisms would require continuous piezometric monitoring beyond the scope of this study. Consequently, rather than generic zoning, we propose a multi-scale “hydro-spatial” governance framework featuring targeted interventions. By establishing strict agricultural redlines in vulnerable overflow zones and implementing eco-hydrological restoration tailored to specific RGT regimes, this paradigm delivers robust methodological insights for advancing precision spatial planning in fragile arid ecosystems. Full article

This study addresses the challenge of detecting white grape bunches (Vitis vinifera L.) in high-density vineyard canopies, a critical task for precision viticulture and yield estimation. Traditional statistical and image-processing methods have struggled to cope with occlusion issues. In this work, more than 200 field RGB images were collected at La Bergonza (Toledo, Spain) and expanded through data augmentation. Several preprocessing strategies were evaluated to enhance bunch visibility. Different convolutional neural network (CNN) architectures were compared, with YOLOv8 outperforming Mask R-CNN in terms of both accuracy and efficiency. YOLOv8, trained for up to 100 epochs on equalized and augmented datasets, achieved outstanding performance, with 84.9% precision, 72.6% recall, and an mAP@0.5 of 83%, far surpassing Mask R-CNN (17% precision and 26% recall). The model successfully detected partially occluded grape bunches, including some that were not visible to human experts, and outperformed previous studies that relied on controlled backgrounds or artificial lighting. The results demonstrate that combining RGB equalization with data augmentation significantly improves detection performance. These findings highlight the potential of deep learning and low-cost RGB imaging systems to enable automated and scalable solutions for yield estimation and canopy analysis. In conclusion, YOLOv8 emerges as a promising tool for accurate grape bunch detection under real field conditions, effectively overcoming previous technological limitations. Full article