Search Results (333)

Precisely identifying and delineating food regions automatically from images, a task known as food image segmentation, is crucial for enabling applications in food science such as automated dietary logging, accurate nutritional analysis, and food safety monitoring. However, accurately segmenting food images, particularly delineating food edges with precision, remains challenging due to the wide variety and diverse forms of food items, frequent inter-food occlusion, and ambiguous boundaries between food and backgrounds or containers. To overcome these challenges, we proposed a novel method called the Multi-view Edge Attention Network (MVEANet), which focuses on enhancing the fine-grained segmentation of food edges. The core idea behind this method is to integrate information obtained from observing food from different perspectives to achieve a more comprehensive understanding of its shape and specifically to strengthen the processing capability for food contour details. Rigorous testing on two large public food image datasets, FoodSeg103 and UEC-FoodPIX Complete, demonstrates that MVEANet surpasses existing state-of-the-art methods in segmentation accuracy, performing exceptionally well in depicting clear and precise food boundaries. This work provides the field of food science with a more accurate and reliable tool for automated food image segmentation, offering strong technical support for the development of more intelligent dietary assessment, nutritional research, and health management systems. Full article

Some bacterial species within the Enterobacteriaceae family possess different types of fimbrial (pili) adhesins that promote adherence to cells and colonization of host tissues. One of the well-characterized fimbrial systems is the pap operon, which encodes P fimbriae, a key virulence factor in urinary and systemic infections. One of the key regulators of P fimbriae is the transcriptional regulator PapB which plays a pivotal role as a master switch, not only by directing phase-variable expression of its own operon but also by influencing expression of heterologous fimbrial systems. This review explores the structural organization, biogenesis, and multi-tiered regulatory control of P fimbriae, with emphasis on PapB and homologous regulatory proteins such as SfaB, FocB, PixB, and PefB. Comparative genomics and phylogenetic analyses reveal that regulators belonging to the PapB family are evolutionarily conserved across π-fimbrial systems and also regulate other types of fimbriae. These regulators respond to epigenetic changes, host-derived signals, and global transcriptional cues to control levels of production of specific fimbriae in a bacterial population to dynamically modulate bacterial adhesion in different environmental niches. Optimally, understanding these mechanisms could lead to novel approaches to perturb PapB-family proteins and abrogate production of some types of fimbriae as a targeted strategy to prevent bacterial infections dependent on adherence mediated by PapB family regulators. Full article

Thermal cameras are becoming popular in several applications of precision agriculture, including crop and soil monitoring, for efficient irrigation scheduling, crop maturity, and yield mapping. Nowadays, these sensors can be integrated as payloads on unmanned aerial vehicles, providing high spatial and temporal resolution, to deeply understand the variability of crop and soil conditions. However, few commercial software programs, such as PIX4D Mapper, can process thermal images, and their functionalities are very limited. This paper reports on the implementation of a custom MATLAB^® R2024a script to extract agronomic information from thermal orthomosaics obtained from images acquired by the DJI Mavic 3T drone. This approach enables us to evaluate the temperature at each point of an orthomosaic, create regions of interest, calculate basic statistics of spatial temperature distribution, and compute the Crop Water Stress Index. In the authors’ opinion, the reported approach can be easily replicated and can serve as a valuable tool for scientists who work with thermal images in the agricultural sector. Full article

Medical imaging is vital for diagnosing and treating colorectal cancer (CRC), a leading cause of mortality. Classifying colorectal polyps and CRC precursors remains challenging due to operator variability and expertise dependence. Deep learning (DL) models show promise in polyp classification but face adoption barriers due to their ‘black box’ nature, limiting interpretability. This study presents an example-based explainable artificial intehlligence (XAI) approach using Pix2Pix to generate synthetic polyp images with controlled size variations and LIME to explain classifier predictions visually. EfficientNet and Vision Transformer (ViT) were trained on datasets of real and synthetic images, achieving strong baseline accuracies of 94% and 96%, respectively. Image quality was assessed using PSNR (18.04), SSIM (0.64), and FID (123.32), while classifier robustness was evaluated across polyp sizes. Results show that Pix2Pix effectively controls image attributes like polyp size despite limitations in visual fidelity. LIME integration revealed classifier vulnerabilities, underscoring the value of complementary XAI techniques. This enhances DL model interpretability and deepens understanding of their behaviour. The findings contribute to developing explainable AI tools for polyp classification and CRC diagnosis. Future work will improve synthetic image quality and refine XAI methodologies for broader clinical use. Full article

Single-view 3D reconstruction remains fundamentally ill-posed, as a single RGB image lacks scale and depth cues, often yielding ambiguous results under occlusion or in texture-poor regions. We propose DP-AMF, a novel Depth-Prior–Guided Adaptive Multi-Modal and Global–Local Fusion framework that integrates high-fidelity depth priors—generated offline by the MARIGOLD diffusion-based estimator and cached to avoid extra training cost—with hierarchical local features from ResNet-32/ResNet-18 and semantic global features from DINO-ViT. A learnable fusion module dynamically adjusts per-channel weights to balance these modalities according to local texture and occlusion, and an implicit signed-distance field decoder reconstructs the final mesh. Extensive experiments on 3D-FRONT and Pix3D demonstrate that DP-AMF reduces Chamfer Distance by 7.64%, increases F-Score by 2.81%, and boosts Normal Consistency by 5.88% compared to strong baselines, while qualitative results show sharper edges and more complete geometry in challenging scenes. DP-AMF achieves these gains without substantially increasing model size or inference time, offering a robust and effective solution for complex single-view reconstruction tasks. Full article

Background/Objectives: The clinically validated multiplex Oncuria bladder cancer (BC) assay quickly and noninvasively identifies disease risk and tracks treatment success by simultaneously profiling 10 protein biomarkers in voided urine samples. Oncuria uses paramagnetic bead-based fluorescence multiplex technology (xMAP^®; Luminex, Austin, TX, USA) to simultaneously measure 10 protein analytes in urine [angiogenin, apolipoprotein E, carbonic anhydrase IX (CA9), interleukin-8, matrix metalloproteinase-9 and -10, alpha-1 anti-trypsin, plasminogen activator inhibitor-1, syndecan-1, and vascular endothelial growth factor]. Methods: In a pilot study (N = 36 subjects; 18 with BC), Oncuria performed essentially identically across three different common analyzers (the laser/flow-based FlexMap 3D and 200 systems, and the LED/image-based MagPix system; Luminex). The current study compared Oncuria performance across instrumentation platforms using a larger study population (N = 181 subjects; 51 with BC). Results: All three analyzers assessed all 10 analytes in identical samples with excellent concordance. The percent coefficient of variation (%CV) in protein concentrations across systems was ≤2.3% for 9/10 analytes, with only CA9 having %CVs > 2.3%. In pairwise correlation plot comparisons between instruments for all 10 biomarkers, R² values were 0.999 for 15/30 comparisons and R² ≥ 0.995 for 27/30 comparisons; CA9 showed the greatest variability (R² = 0.948–0.970). Standard curve slopes were statistically indistinguishable for all 10 biomarkers across analyzers. Conclusions: The Oncuria BC assay generates comprehensive urinary protein signatures useful for assisting BC diagnosis, predicting treatment response, and tracking disease progression and recurrence. The equivalent performance of the multiplex BC assay using three popular analyzers rationalizes test adoption by CLIA (Clinical Laboratory Improvement Amendments) clinical and research laboratories. Full article

Ulmus glabra Huds. is a mesophilic, montane broadleaf tree with high ecological value, commonly found in temperate riparian and floodplain forests across Türkiye. Its populations in Türkiye have declined due to anthropogenic disturbances and climatic pressures that cause habitat fragmentation and threaten the species’ long-term survival. In this research, we used Maximum Entropy (MaxEnt) to build species distribution models (SDMs) and applied the Restoration Planner (RP) tool to identify and prioritize critical restoration sites under both current and projected climate scenarios (SSP245, SSP370, SSP585). The SDMs highlighted areas of high suitability, primarily along the Black Sea coast. Future projections show that habitat fragmentation and shifts in suitable areas are expected to worsen. To systematically compare restoration options across different future scenarios, we derived and applied four spatial network status indicators using the RP tool. Specifically, we calculated Restoration Pixels (REST_PIX), Average Distance of Restoration Pixels from the Network (AVDIST_RP), Change in Equivalent Connected Area (ΔECA), and Restoration Efficiency (EFFIC) using the RP tool. For the 1 <-> 2 restoration pathways, the highest efficiency (EFFIC = 38.17) was recorded under present climate conditions. However, the largest improvement in connectivity (ΔECA = 60,775.62) was found in the 4 <-> 5 pathway under the SSP585 scenario, though this required substantial restoration effort (REST_PIX = 385). Temporal analysis noted that the restoration action will have most effectiveness between 2040 and 2080, while between 2081 and 2100, increased habitat fragmentation can severely undermine ecological connectivity. The result indicates that incorporation of habitat suitability modeling into restoration planning can help to design cost-effective restoration actions for degraded land. Moreover, the approach used herein provides a reproducible framework for the enhancement of species sustainability and habitat connectivity under varying climate conditions. Full article

Background/Objectives: To evaluate the concentrations of trace elements in tear fluid among athletes using particle-induced X-ray emission (PIXE), and to assess the associations with gender, sports intensity, and nutritional supplement intake. Methods: In this cohort study, 84 athletes engaged in high- or low-intensity sports completed a demographic and supplement-use questionnaire. Tear samples were collected using Schirmer strips and analyzed for elemental composition with PIXE, a high-sensitivity technique suited for small biological samples. Multivariate and nonparametric statistical analyses were used to compare groups. Results: There were 46 males and 38 females, aged 17–63 years (mean 30.21 years). Tear phosphorus, potassium, and sulfur concentrations were higher in women than men and higher in women participating in low-intensity compared to high-intensity sports. Tear concentrations of magnesium were higher in men participating in high-intensity sports compared to low-intensity sports. They were higher in men than women regardless of supplement intake. Iron concentrations were higher in men than women only when neither group was taking supplements. Smoking had a slight inverse relationship to iron values. Iron levels were particularly high in men participating in intense sports and low in smokers. Magnesium supplements were associated with raised magnesium levels in tears. Conclusions: This study demonstrates an association between trace element levels in human tears and gender, sports intensity, and food supplement intake. PIXE enables the evaluation of trace element concentration in tears, which may serve as potential biomarkers for the clinical assessment of athletes’ health. Full article

Pathological analysis of tissue biopsies remains the gold standard for diagnosing cancer and other diseases. However, this is a time-intensive process that demands extensive training and expertise. Despite its importance, it is often subjective and not entirely error-free. Over the past decade, pathology has undergone two major transformations. First, the rise in whole slide imaging has enabled work in front of a computer screen and the integration of image processing tools to enhance diagnostics. Second, the rapid evolution of Artificial Intelligence has revolutionized numerous fields and has had a remarkable impact on humanity. The synergy of these two has paved the way for groundbreaking research aiming for advancements in digital pathology. Despite encouraging research outcomes, AI-based tools have yet to be actively incorporated into therapeutic protocols. This is primary due to the need for high reliability in medical therapy, necessitating a new approach that ensures greater robustness. Another approach for improving pathological diagnosis involves advanced optical methods such as spectral imaging, which reveals information from the tissue that is beyond human vision. We have recently developed a unique rapid spectral imaging system capable of scanning pathological slides, delivering a wealth of critical diagnostic information. Here, we present a novel application of spectral imaging (SI) for virtual Hematoxylin and Eosin (H&E) staining using a custom-built, rapid Fourier-based SI system. Unstained human biopsy samples are scanned, and a Pix2Pix-based neural network generates realistic H&E-equivalent images. Additionally, we applied Principal Component Analysis (PCA) to the spectral information to examine the effect of down sampling the data on the virtual staining process. To assess model performance, we trained and tested models using full spectral data, RGB, and PCA-reduced spectral inputs. The results demonstrate that PCA-reduced data preserved essential image features while enhancing statistical image quality, as indicated by FID and KID scores, and reducing computational complexity. These findings highlight the potential of integrating SI and AI to enable efficient, accurate, and stain-free digital pathology. Full article

A series (n = 22) of glasses from the site Kapucinski vrt (garden of the Capuchin monastery, 5th–17th c. CE) in Koper (Capodistria), a port town in the northern Adriatic, was measured using a combined PIXE and PIGE method. Koper has been continuously populated since the late Roman period, with a rich medieval history, thus offering an opportunity to study Early Medieval glass. Stemmed goblet fragments, in the original publication dated between the 6th–9th centuries CE, and several other vessel types (beakers and flasks or bottles and lamps) were selected for analysis. The measurements were expected to show the trends in glass production and consumption from Late Antiquity until the Middle Ages, notably the transition between natron to plant ash glass and the supply of fresh glass. Among the set of 22 glass vessel fragments, both natron and plant ash glass were identified. For finer classification, we relied on a newly developed method of Euclidean distances with respect to major concentrations. Natron glass of the types Foy 2.1 (9 examples), Magby (2 examples), and Levantine I (Apollonia; 2 examples) was found. Two glasses remain undetermined but testify to an Egyptian origin. Most natron glasses show signs of recycling. Among the three unrecycled glasses (about 20% of the whole set), there are two examples of Levantine glass and a Magby glass lamp; this may indicate a modest supply of fresh glass during the period. Plant ash glass may be attributed to the Early or High Middle Ages, exploiting the purified alkalis of the Levantine coasts (known as alume catino in later Venetian glassmaking), and the admixture of impurities in the siliceous sands suggests the circulation and consumption of glass that was produced and traded in the eastern Mediterranean since the 10th century CE. Full article

The growing interest in weight management has elevated the popularity of liposuction. Individuals deciding whether to undergo liposuction must rely on a doctor’s subjective projections or surgical outcomes for other people to gauge how their own body shape will change. However, such predictions may not be accurate. Although deep learning technology has recently achieved breakthroughs in analyzing medical images and rendering diagnoses, predicting surgical outcomes based on medical images outside clinical settings remains challenging. Hence, this study aimed to develop a method for predicting body shape changes after liposuction using only images of the subject’s own body. To achieve this, we utilize data augmentation based on a conditional continuous Generative Adversarial Network (CcGAN), which generates realistic synthetic data conditioned on continuous variables. Additionally, we modify the loss function of Pix2Pix—a supervised image-to-image translation technique based on Generative Adversarial Networks (GANs)—to enhance prediction quality. Our approach quantitatively and qualitatively demonstrates that accurate, intuitive predictions before liposuction are possible. Full article

Many image-to-image translation tasks face an inherent problem of asymmetry in the domains, meaning that one of the domains is scarce—i.e., it contains significantly less available training data in comparison to the other domain. There are only a few methods proposed in the literature that tackle the problem of training a CycleGAN in such an environment. In this paper, we propose a novel method that utilizes pdf (probability density function) distance-based augmentation of the discriminator network corresponding to the scarce domain. Namely, the method involves adding examples translated from the non-scarce domain into the pool of the discriminator corresponding to the scarce domain, but only those examples for which the assumed Gaussian pdf in VGG19 net feature space is sufficiently close to the GMM pdf that represents the relevant initial pool in the same feature space. In experiments on several datasets, the proposed method showed significantly improved characteristics in comparison with a standard unsupervised CycleGAN, as well as with Bootstraped SSL CycleGAN, where translated examples are added to the pool of the discriminator corresponding to the scarce domain, without any discrimination. Moreover, in the considered scarce scenarios, it also shows competitive results in comparison to fully supervised image-to-image translation based on the pix2pix method. Full article

Remotely Piloted Aircraft (RPA) as sensor-carrying airborne platforms for indirect measurement of plant physical parameters has been discussed in the scientific community. The utilization of RGB sensors with photogrammetric data processing based on Structure-from-Motion (SfM) and Light Detection and Ranging (LiDAR) sensors for point cloud construction are applicable in this context and can yield high-quality results. In this sense, this study aimed to compare coffee plant height data obtained from RGB/SfM and LiDAR point clouds and to estimate soil compaction through penetration resistance in a coffee plantation located in Minas Gerais, Brazil. A Matrice 300 RTK RPA equipped with a Zenmuse L1 sensor was used, with RGB data processed in PIX4D software (version 4.5.6) and LiDAR data in DJI Terra software (version V4.4.6). Canopy Height Model (CHM) analysis and cross-sectional profile, together with correlation and statistical difference studies between the height data from the two sensors, were conducted to evaluate the RGB sensor’s capability to estimate coffee plant height compared to LiDAR data considered as reference. Based on the height data obtained by the two sensors, soil compaction in the coffee plantation was estimated through soil penetration resistance. The results demonstrated that both sensors provided dense point clouds from which plant height (R2 = 0.72, R = 0.85, and RMSE = 0.44) and soil penetration resistance (R2 = 0.87, R = 0.8346, and RMSE = 0.14 m) were accurately estimated, with no statistically significant differences determined between the analyzed sensor data. It is concluded, therefore, that the use of remote sensing technologies can be employed for accurate estimation of coffee plantation heights and soil compaction, emphasizing a potential pathway for reducing laborious manual field measurements. Full article

Contrastive image representation learning has been essential for pre-training vision foundation models to deliver excellent transfer learning performance. It was originally developed based on instance discrimination, which focuses on instance-level recognition tasks. Lately, the focus has shifted to directly working on the dense spatial features to improve transfer performance on dense prediction tasks such as object detection and semantic segmentation, for which pixel-level and region-level contrastive learning methods have been proposed. Region-level methods usually employ region-mining algorithms to capture holistic regional semantics and address the issue of semantically inconsistent scene image crops, as they assume that pixel-level learning struggles with both. In this paper, we revisit pixel-level learning’s potential and show that (1) it can effectively and more efficiently learn holistic regional semantics and (2) it intrinsically provides tools to mitigate the impact of semantically inconsistent views involved with scene-level training images. We prove this by proposing PixCon, a pixel-level contrastive learning framework, and testing different positive matching strategies based on this framework to rediscover the potential of pixel-level learning. Additionally, we propose a novel semantic reweighting approach tailored for pixel-level learning-based scene image pre-training, which outperforms or matches the performance of previous region-level methods in object detection and semantic segmentation tasks across multiple benchmarks. Full article

Multi-view 3D reconstruction aims to discover 3D characteristics based on visual information captured across multiple viewpoints. Transformer networks have shown remarkable success in various computer vision tasks, including multi-view 3D reconstruction. However, the reconstruction of accurate 3D shapes faces challenges when trying to efficiently extract and merge features across views. The existing frameworks struggled to capture the subtle relationships between the views, resulting in a poor reconstruction. To address this issue, we present a new framework, WGA-SWIN, for 3D reconstruction using multi-view objects. Our method introduces a Window Grouping Attention (WGA) mechanism that uses group tokens from different views for each window attention operation, enabling efficient inter-view and intra-view feature extraction. Diversity among various groups in a model contributes to the richness of feature learning, which results in advanced and dependable feature learning, resulting in more comprehensive and robust representations. Within the encoder swin transformer blocks, we integrated WGA to utilize both hierarchical design and shifted window attention mechanisms for efficient multi-view feature extraction. In addition, we developed a progressive hierarchical decoder that combines swin transformer blocks with 3D convolutions to utilize voxel representation, resulting in a high resolution for obtaining high-quality 3D reconstructions with fine structural details. The experimental results on the benchmark datasets ShapeNet and Pix3D demonstrate that our work achieves state-of-the-art (SOTA) performance, outperforming existing methods in both single-view and multi-view 3D reconstruction, beyond the capabilities of current technologies. We lead by 0.95% and 1.07% in both IoU and F-Scores respectively, which demonstrates the robustness of our method. Full article