Search Results (166)

Magnetic anomaly detection is vulnerable to environmental noise and insufficient prior target information, making non-periodic anomaly signals difficult to detect at low-signal-to-noise-ratio (SNR) conditions. This paper proposes a detection method based on a multi-parameter-constrained mirror dual-branch biased monostable stochastic resonance (SR) system. Nonlinear odd-order bias terms are introduced into the conventional biased monostable potential function to build a multi-parameter-controllable SR model. This improves regulation of potential-well width, depth, and wall morphology, enhancing noise-energy utilization and responses to non-periodic features. Considering peak-type, valley-type, and bipolar anomaly morphologies, a mirror dual-branch SR structure is developed to cooperatively detect features with different polarities. To preserve temporal waveforms and time–frequency structures during parameter optimization, a composite metric combining the correlation coefficient and wavelet-domain image structural similarity index is constructed. Multi-fidelity robust Bayesian optimization is used to obtain a unified robust parameter set for the magnetic anomaly signal family. Experiments with simulated colored noise and measured geomagnetic noise show that the proposed method effectively recovers magnetic anomaly features under strong noise. At −19 dB SNR, its detection probability remains above 80%. Compared with orthogonal basis function decomposition, empirical mode decomposition, and complete ensemble empirical mode decomposition with adaptive noise, the method achieves better noise suppression, feature preservation, and detection performance under low-SNR conditions. Full article

Referring Expression Comprehension (REC) localizes a target object in an image given a natural-language referring expression and is a core benchmark for fine-grained vision–language alignment. Recent detection-style multimodal Transformers achieve strong REC performance but typically rely on high-capacity visual and textual backbones, incurring substantial storage and compute costs. Replacing these backbones with lightweight alternatives greatly reduces model size, yet often degrades cross-modal alignment and yields a persistent accuracy gap. We propose CARM, a minimally invasive Cross-modal Alignment Recovery Module inserted between lightweight backbones and the downstream multimodal Transformer, requiring no changes to either component. CARM injects complementary priors via bidirectional prompts and uses a Cross-Attention Gate (CAG) to adaptively filter and scale prompt-induced updates; it further integrates Tree-of-Attributes Prompts (TAPs) to strengthen fine-grained grounding of attributes such as color, size, and spatial location. On RefCOCO, switching to lightweight backbones drops P@1 (IoU ≥ 0.5) to 84.51, while CARM improves it to 86.23, recovering most of the loss. Meanwhile, the overall model storage (checkpoint) is reduced by about 80%, demonstrating that the cross-modal alignment degradation induced by compression can be effectively restored. Full article

Underwater images often suffer from color cast, low contrast, scattering-induced haze, and texture degradation, which limit the performance of underwater visual perception systems. To address these problems, this study proposes FreqMambaGAN, a frequency-decoupled selective state-space cycle-adversarial network for underwater image enhancement. The proposed method is built upon a CycleGAN-style bidirectional translation framework and introduces a frequency-decoupled Mamba generator to separately model low-frequency color and illumination information and high-frequency texture and edge details. In addition, Efficient Mamba Blocks are embedded into the generator and discriminator to enhance long-range dependency modeling with linear computational complexity. Skip-attention connections are further adopted to preserve shallow spatial details during reconstruction. To improve training stability and imaging plausibility, a multi-stage training strategy is designed by combining supervised warm-up, unpaired cycle-adversarial learning, perceptual regularization, total variation smoothing, and a lightweight physics-inspired consistency constraint based on dark-channel and underwater image-formation priors. Experiments on public underwater image enhancement datasets demonstrate that FreqMambaGAN achieves competitive quantitative performance and visually improved enhancement results in terms of color correction, contrast restoration, haze suppression, and structural preservation. These results indicate that integrating frequency-domain decomposition with selective state-space modeling is effective for underwater image enhancement. Full article

This study advances the model of proactive caring to articulate the strategies and practices employed by higher education professionals to support the retention and graduation of underserved students. Employing an explanatory case study design and drawing upon multiple data sources—including semi-structured interviews, focus groups, and observations—collected from students and higher education professionals, we identified six core elements essential to proactive caring: (1) staff accessibility, (2) trust-building, (3) positive motivation, (4) reinforcement, (5) encouragement, and (6) student accountability. Our findings also reveal that higher educational professionals initiate support prior to students’ arrival on campus by assessing institutional environments to identify potential barriers and leveraging data and experiential knowledge to proactively address these challenges. This research contributes to ongoing efforts to promote access and equity in higher education, particularly in the context of increasing anti-diversity, equity, and inclusion (DEI) legislation. While the current study centers on Students of Color in STEM disciplines, the elements underlying the proactive caring model may be broadly applicable, which offers a framework for empathetic practitioners committed to supporting underserved students throughout their collegiate journey. Full article

The present study evaluates the optical and colorimetric properties of Polylactic acid (PLA)-based films blended with Poly(3-hydroxybutyrate) (PHB) and Poly(butylene succinate) (PBS) and impregnated with mango leaf extract (MLE) using supercritical solvent impregnation (SSI) under different operating conditions (pressure: 10–30 MPa; temperature: 35–55 °C). Additionally, the relationship between impregnation load (IL) and color properties, as well as the release kinetics of the impregnated compounds, was investigated. The incorporation of PHB and PBS into the PLA matrix prior to impregnation led to a slight increase in the b* parameter (from 1.64 to 2.61), indicating a tendency toward yellowish tones. After SSI, all films exhibited noticeable color changes, with a shift toward yellowish-green hues and a decrease in lightness, regardless of processing conditions. Statistical analysis confirmed that polymer composition and its interaction with pressure and temperature significantly affected color properties (p-value < 0.001). The addition of PHB and PBS, as well as MLE impregnation, enhanced UV-barrier properties, while also modifying film transparency and opacity. In particular, PLA-PBS films showed higher opacity (more than 20 times) and lower transparency compared to neat PLA. These films also exhibited the highest IL values (2.41–4.75 mg MLE/100 mg polymer). Multivariate regression analysis demonstrated a strong correlation between CIELAB parameters (L*, a*, and b*) and IL (R² > 85%, p-value < 0.001). Release studies in a food simulant showed partial release profiles, well described by Peleg’s model (R² > 0.90). Furthermore, Korsmeyer–Peppas model fitting yielded diffusion exponents (n < 0.5), indicating quasi-Fickian diffusion mechanisms governing the release process. Full article

Weak underwater acoustic signal detection is fundamentally challenged by low signal-to-noise ratio (SNR), colored ocean noise, multipath distortion, and environmental mismatch. Existing weak-signal detectors have mainly focused on spectral enhancement, time-frequency tracking, or fixed-environment model matching, while environmentally robust Bayesian methods have been developed primarily for localization, matched-field processing, and channel estimation rather than weak passive detection itself. To bridge this gap, this paper proposes a hierarchical Bayesian detector for weak underwater acoustic signal detection under environmental mismatch. The received observation is modeled by jointly incorporating structured weak-signal coefficients, target-related parameters, and uncertain environmental parameters into a unified Bayesian hypothesis-testing framework. In particular, the acoustic environment is treated as a latent random variable rather than a fixed nominal condition so that robustness can be achieved through environmental marginalization. Since the resulting marginal likelihood is analytically intractable, a variational Bayesian approximation is developed to derive a tractable evidence-based detection statistic. Numerical simulations under low-SNR, multipath-distorted, and environmentally uncertain underwater conditions demonstrate that the proposed detector achieves consistently strong performance under both matched and mismatched scenarios. Ablation results in controlled simulations further indicate that environmental marginalization provides the largest observed robustness gain, whereas the structured weak-signal prior offers an additional improvement in weak-signal discrimination. These results provide controlled simulation-based evidence for the potential of hierarchical Bayesian inference in robust passive underwater acoustic detection under prescribed environmental uncertainty models. Full article

Automated and objective grading of cocoa (Theobroma cacao L.) fermentation remains a major challenge because the conventional cut test relies on subjective visual inspection and is difficult to scale. In this study, we develop and evaluate a deep learning pipeline for classifying cocoa bean fermentation levels from expert-annotated cut-test images acquired under controlled conditions, enabling the systematic evaluation and comparison of multiple convolutional and transformer-based architectures under consistent preprocessing, training, and evaluation protocols. The dataset comprises 4347 segmented cocoa bean images distributed across four severely imbalanced classes, namely fermented, under-fermented, slaty, and violet. Representative architectures, including EfficientNet-B0, MobileNetV3-Large, ConvNeXt-XLarge, ViT-Base, and ViT-Large, are benchmarked to analyze the effects of class imbalance, RGB versus HSV color representation, training duration, and label-space formulation. The results show that severe class imbalance strongly degrades performance in direct four-class classification. A hierarchical binary-to-multiclass strategy significantly improves balanced recognition by separating fermented from unfermented beans prior to subclass discrimination, increasing macro-F1 scores from approximately 80–83% to 89–91%. Among the evaluated models, ViT-Base emerges as the most stable architecture across experimental settings and offers the best balance between classification performance, training stability, and computational cost. Although larger models achieve slightly higher peak performance under balanced conditions, ViT-Base provides more consistent results under realistic constraints. The proposed framework enables near-real-time inference on segmented single-bean images and supports objective, reproducible, and scalable fermentation assessment. These findings demonstrate that performance in cocoa fermentation grading is determined not only by model capacity, but also by imbalance-aware label-space design and evaluation protocols aligned with real-world cut-test conditions. Full article

Accurate estimation of atmospheric light (AL) is essential within the atmospheric scattering model (ASM) to achieve high-quality image dehazing. Most existing methods, however, typically assume spatial uniformity of AL and rely on heuristic estimation from distant pixels, which often results in color distortion and exposure imbalance in dehazed outputs. To address this issue, we propose a novel framework that decouples AL into distinct color and intensity components. Specifically, a neutral pixel prior (NPP) is introduced for precise AL color estimation, which can eliminate color casts. For AL intensity estimation, an adaptive global-local fusion strategy integrating luminance perception transformation and a depth-related color prior (DRCP) is developed to realize balanced exposure. Extensive experiments demonstrate that the proposed method significantly outperforms state-of-the-art AL estimation methods, yielding dehazed images with enhanced color fidelity and more natural illumination. Full article

Underwater image enhancement is essential for marine visual perception tasks. However, the highly heterogeneous optical degradations in real-world waters, the scarcity of paired training data, and the inherent dilemma for existing models in balancing long-range dependency modeling with computational overhead pose significant challenges. To address these issues, this paper proposes a prior-guided, self-supervised underwater image enhancement framework called Scene-Aware Retinex Mamba (SARM). This framework seamlessly integrates Retinex theoretical priors with state space models (SSMs) and operates without paired supervision by employing a prior-guided pseudo-labeling strategy to guide network optimization. Architecturally, SARM deeply couples the physical Retinex prior with SSM. Its core module integrates multi-color space features and leverages a 2D selective scan mechanism to achieve global context modeling with linear complexity $\mathcal{O}\left(HW\right)$, effectively removing complex color casts and suppressing non-uniform scattering noise. To further overcome the generalization bottlenecks in cross-domain underwater testing, this paper introduces a Scene-Aware Adapter (SAA), which facilitates dynamic loss scheduling and adaptive feature gating by quantifying scene-specific degradation characteristics. Comprehensive evaluations on multiple benchmark datasets, including UIEB, EUVP, and UCCS, demonstrate that SARM achieves state-of-the-art subjective and objective enhancement quality (e.g., yielding a URanker score of 2.491 and a CCF score of 35.76), while maintaining an ultra-fast inference speed of 136.52 FPS on the UIEB dataset. Furthermore, extended experiments reveal that SARM can significantly boost the performance of downstream vision tasks, validating its potential as a robust preprocessing module for various practical marine vision applications. Full article

Hot air drying of pumpkin (Cucurbita moschata) and its by-products, mainly the peel and placenta with seeds, has been investigated, analysing the influence of pretreatments on drying kinetics and bioactive compound content. Pumpkin flours not only stand out for their microbiological stability (low water activity), but also for their bioactive compounds important for health, including phenolic compounds and other antioxidants. Pretreatments prior to drying may improve both the quality and the drying efficiency, although their optimization has not been studied in pumpkin by-product flours. Hence, we studied different pretreatments, such as steam blanching (SB) and freezing (F) (−18 °C), to investigate their effect on the pumpkin by-product flour quality (color and water activity) after drying and compared to flours made with the edible part (pulp). In addition, different drying kinetics models were evaluated. SB pulp and peel achieved lower water activity (0.25/0.20) than F (0.35/0.36), compared to untreated pumpkin by-product flour (0.40/0.45). The SB placenta with seeds achieved a lower water activity (0.19) than F (0.55). The total phenolic content (TPC) increased up to 41.7, 60.2 and 40.9% in pre-treated and dried pulp, peel and placenta with seeds, respectively, compared to control (CTRL). A similar result was obtained for total flavonoid content (TFC), where increases of up to 86.4, 36.4, and 32.2% were observed in pre-treated pulp, peel and placenta with seeds, respectively. For total antioxidant capacity (TAC), results showed an increase of up to 33.1% in pre-treated peel, with slight reductions in pulp (20.1%) and placenta with seeds (39.1%), compared to CTRL. These results are promising for obtaining new dehydrated plant products from industrial processing by-products, while maintaining their beneficial health characteristics. These powdered products can be used in future research on the formulation of new fortified foods, such as bakery products, or the development of natural additives for beverages or soups. Full article

Underwater salient instance segmentation (USIS) remains challenging because color distortion, low contrast, and scattering often weaken appearance cues, while reliable geometric measurements are usually unavailable. Existing methods mainly rely on red–green–blue (RGB) information, which can be insufficient in visually degraded underwater scenes. We propose the Depth-guided Cross-modal Fusion Network (DCFNet), a depth-guided fusion framework that leverages pseudo-depth estimated from RGB images as an auxiliary structural prior. DCFNet contains a dual-branch encoder, a cross-modal fusion branch, a refinement decoder, and an instance branch. In the fusion branch, the proposed Depth-Aware Modality Injection (DAMI) module selectively exchanges information between RGB and pseudo-depth features to reduce the influence of noisy depth estimates. The decoder further combines Inverted Residual Transformer (IRT) blocks and Bidirectional Attention Gate (BiAG) modules for contextual modeling and boundary refinement. Finally, the instance branch integrates positional cues to generate dynamic kernels for proposal-free mask prediction. Experiments on USIS10K and USIS16K show that DCFNet achieves competitive performance against several relevant baselines. Ablation studies further indicate that both the pseudo-depth prior and the proposed fusion architecture contribute to the final performance. Full article

Multidimensional line spectral estimation plays a fundamental role in communication and sensing systems, where it is often used for estimating channel parameters such as angles of arrival and time delays. Existing channel parameter estimation methods often suffer from limited resolution, high computational complexity, or strong sensitivity to noise, and the multidimensional variational line spectral estimation (MDVALSE) algorithm, although effective in off-grid estimation, degrades significantly under low signal-to-noise ratio (SNR) conditions. Recently, generative models, especially diffusion models, have demonstrated strong capabilities in prior-guided denoising and reconstruction of noise-contaminated signals by effectively learning the underlying data structure. Motivated by this, we propose a diffusion-enhanced multidimensional variational line spectral estimation algorithm for channel parameter extraction. Specifically, a diffusion model is first employed to denoise the estimated channel response and improve the observation quality. Then, considering that the residual error after diffusion-based denoising is generally colored rather than white, a colored-noise extension of MDVALSE, termed C-MDVALSE, is derived to better match the statistical structure of the denoised observations. Simulation results in various scenarios show that the proposed algorithm achieves more accurate channel reconstruction and channel parameter estimation than MDVALSE and other existing methods, with particularly significant improvements in low-SNR regimes. Full article

DNNs are highly vulnerable to adversarial examples (AEs). To achieve high transferability, traditional AEs often introduce unnatural artifacts that are easily perceptible to the human eye. Unrestricted attacks have emerged as a promising paradigm to generate more natural unrestricted adversarial examples (UAEs). However, existing UAEs struggle to balance visual fidelity and black-box transferability. Color-based attacks produce noticeable unnatural visual mutations, and diffusion-based attacks transfer poorly to unknown black-box models. We observe that directly injecting unconstrained random perturbations into the diffusion latent space destroys the normal distribution of data, thereby causing a distribution shift. Distribution shifts degrade adversarial perturbations into invalid noise and cause surrogate model overfitting. Furthermore, introducing elastic deformation during the denoising process forces surrogate models to focus on highly transferable features. As a result, we propose an unrestricted attack based on deformation-constrained diffusion, called DeDiAttack. Our method utilizes the manifold prior knowledge of diffusion models to translate elastic deformations into smooth fluid changes. The mechanism effectively eliminates unnatural artifacts and generates highly natural and transferable UAEs. Extensive black-box experiments demonstrate that DeDiAttack outperforms existing attacks and improves the black-box transferability of generated UAEs by 7.2% on the ViT-B surrogate model. The proposed method also provides a useful robustness evaluation tool for vision-based sensing and imaging systems. Full article

Accurate high-throughput fruit detection is the core prerequisite for precision citrus management. Existing models face a critical trade-off between accuracy for small fruits and computational efficiency, restricting large-scale industry transformation. To resolve this, we propose Citrus-SGYOLOv2, an optimized deep learning architecture specifically engineered for high-throughput phenotypic monitoring. The primary contribution of this work lies in three synergistic innovations: a novel Symmetric Ghost Backbone that prunes architectural redundancy while maintaining hierarchical feature depth; a Citrus Color Prior Calibration Attention Mechanism (Citrus_SE) that embeds physiological chromaticity priors to suppress complex spectral noise from foliage; and a P2-layer-based full-scale fusion strategy designed to recover fine-grained spatial details lost during downsampling. Experiments on our self-built dataset show that Citrus-SGYOLOv2 achieves 95.54% mAP@50 and 77.13% mAP@50–95, outperforming YOLOv11s by 5.03 and 9.90 percentage points respectively. Notably, the model achieves a 48.8% reduction in parameters (4.84 M) while sustaining a high-throughput inference speed of 139.00 FPS. This research provides a robust and efficient foundational framework for intelligent yield estimation and precision orchard management. Full article

Background: The characterization of thyroid lesions is essential in clinical practice. Recent advances in imaging modalities, including nuclear imaging (NM), color Doppler ultrasonography, and sonography, have markedly improved the diagnostic accuracy for thyroid nodules. Objective: To assess thyroid diseases using Doppler ultrasound, nuclear scintigraphy, and sonography. Results: In this cross-sectional single-center study, 144 patients were examined to determine their thyroid structure and function using a multimodal imaging approach. Fine-needle aspiration cytology (FNAC) indicated that most thyroid nodules were benign (62.5%), with 37.5% being malignant. Doppler vascularity demonstrated a sensitivity of 70.4% and a specificity of 40% (AUC = 0.514) for malignancy detection, while scintigraphy uptake in hypofunctioning nodules (nodules with decreased radionuclide uptake) showed a sensitivity of 37% and a specificity of 54.4% (AUC = 0.388). Thyroxine hormone levels showed a sensitivity of 57.4% and a specificity of 45.6% (AUC = 0.503) for detecting malignant thyroid nodules. In multivariate logistic regression, increased Doppler vascularity remained an independent predictor of malignancy (OR = 2.39; 95% CI: 1.15–4.96; p = 0.019), whereas decreased scintigraphic uptake showed a borderline effect (OR = 1.82; p = 0.069); high T4 level and increased uptake were not significant predictors. The combined Doppler ultrasound, scintigraphy, and thyroxine level model yielded an AUC of 0.72 (95% CI: 0.63–0.81), markedly higher than any single parameter. At the optimal Youden threshold (0.43), the model achieved 79.6% sensitivity, 68.2% specificity, and 72.4% accuracy, highlighting the superior diagnostic performance of the integrated approach for pre-FNAC stratification of thyroid malignancies. There was a strong, significant linear association between thyroxine levels and thyroid scintigraphy uptake (p-value < 0.001). Most patients with normal thyroxine levels exhibited decreased uptake (66.1%), whereas a minority (6.5%) demonstrated elevated uptake levels. This study found a strong correlation between mixed-echogenicity nodules and thyroid scintigraphy uptake (p-value = 0.019). Mixed-echogenicity nodules were most often associated with reduced uptake (57.8%), and hypoechoic nodules often had normal uptake (57.1%). Conclusions: The complementary integration of color Doppler vascularity, Tc-99m thyroid scintigraphy, and serum thyroxine levels yields superior Doppler–scintigraphy uptake correlation, increases the overall diagnostic accuracy, and offers a practical, non-invasive algorithm for differentiating benign from malignant thyroid nodules prior to FNAC or surgery. Full article