Search Results (42)

Detecting small anomalous targets in hyperspectral imagery is challenging when neither target spectra nor spatial morphologies are known, because most unsupervised detectors still encode implicit assumptions about background statistics or target structures and therefore generalize poorly to genuinely unknown threats. This paper introduces the Hard-Negative-Guided Topology Transfer Network (HNGT-Net), a lightweight teacher–student framework that addresses this problem from two perspectives. On the representation side, a similarity graph coupling spatial adjacency with feature-space nearest neighbors characterizes the spectral–spatial topology of normal backgrounds, and a three-level consistency objective over node embeddings, structural relations, and graph-Laplacian responses transfers this topology from a frozen teacher to a compact student. On the discrimination side, hard negatives are synthesized directly on pure normal pixels through sparsity-gated projected gradient perturbation, while a response-margin constraint forces the student to score such negatives above normal samples and mitigates distribution overfitting. Experiments on four public benchmark scenes and one Salinas-derived synthetic dataset show that HNGT-Net achieves an average AUC of 0.9953 with the smallest variance among all competitors. The student branch provides 10.5-fold parameter compression relative to the teacher, while the complete teacher–student stack remains compact, supporting resource-constrained remote sensing deployment. Full article

Product concept images combine symmetry, closure, balance, and repeated components. Existing shape analysis and computational aesthetic methods can quantify these properties; however, when evidence is reduced to global descriptors or aggregate scores, image-layer provenance and sensitivity to rasterization or heuristic settings may be obscured. This paper presents a traceable image-processing pipeline based on scenario framing, alternative specification, geometry-informed computation, evidence synthesis, and design embodiment (SAGE-D), evaluated on four controlled smart-speaker alternatives using separate body, light-band, and aperture masks. Seven dimensionless descriptors measure silhouette reflection, centroid balance, light-band closure, aperture regularity and gradient, component-scale retention, and contour compactness. Resolution resampling, one-pixel morphology, parameter perturbation, and synthetic controls assess sensitivity. At $512\times 512$ pixels, A, B, and R showed exact bilateral silhouette consistency; B and R showed complete light-band occupancy; and C and R showed strong downward aperture-radius gradients. Conventional same-mask measures gave concordant geometric readings, while leave-one-gate-out analysis showed that screening depended mainly on predefined closed-ring and linear-gradient requirements. Only R passed all six case gates. SAGE-D is used here as an auditable organization of layer-specific measurements and bounded screening rules, not as a superior descriptor set. The conclusions are limited to the supplied two-dimensional (2D) representations and do not establish population-level generalizability, preference, or engineering performance. Full article

Permafrost maps are important for regional planning in cold regions, but binary products do not represent gradients in thermal conditions, seasonal thaw response, or mapping confidence. This study developed an uncertainty-aware regional permafrost-condition screening framework for the Daxing’anling region during 2003–2022 by integrating TTOP-derived mean annual ground temperature (MAGT), reconstructed permafrost-occurrence probability, and Kudryavtsev-model-derived active-layer thickness (ALT). MAGT ranged from −4.11 to 5.28 °C, with a mean of −0.28 °C, and comparison with 23 published borehole records from 17 reported locations yielded r = 0.756 and RMSE = 0.419 °C. Explicit measurement depths were available for 12 of the 23 records and ranged from 10 to 15 m. Model-derived ALT ranged from 1.480 to 1.864 m. The complete point-scale evaluation using all 12 valid maximum depth of seasonal thaw (MDST) records yielded an RMSE of 0.92 m. Adding ALT changed 21.05% of valid-pixel assignments. Cold–low-response permafrost, Cold–moderate-response permafrost, Warm–enhanced-response permafrost, Near-thaw transitional permafrost, Marginal/low-confidence permafrost, and Non-permafrost occupied 7.3%, 17.1%, 14.9%, 4.1%, 13.9%, and 42.8% of the domain, respectively. Full Monte Carlo uncertainty propagation retained 78.65% modal agreement with the deterministic baseline, and 56.65% of the domain had a maximum class-membership probability below 0.60, whereas threshold-only perturbation retained 98.92% agreement. The framework is therefore suited to regional investigation and monitoring prioritization; project-level engineering decisions require direct geotechnical and deformation-based evidence. Full article

Although coronary CT angiography (CCTA) is widely utilized for diagnosing coronary artery disease (CAD), automated CCTA image segmentation is frequently hindered by sparse annotations, pseudo-label noise, and under-delineated fine branches. To mitigate these issues, we present PCDW-Net, a semi-supervised segmentation framework that couples perturbation consistency with discrepancy-aware weighting for enhanced label-scarce performance. Utilizing Adaptive Multi-scale Attention Fusion Network (AMAF-Net) as the backbone within a teacher-student architecture, the network applies diverse perturbations to unlabeled samples, leveraging a consistency loss to promote feature invariance. Simultaneously, a pixel-level discrepancy-aware weighting scheme serves to suppress erroneous pseudo-labels. Evaluated on the public ASOCA and private CTA40 datasets using 10% and 20% annotated fractions, the model was evaluated using Dice similarity coefficient (DSC) and Average Symmetric Surface Distance (ASSD). Under the 20% labeling constraint, PCDW-Net yielded a DSC of 85.36% on ASOCA and 83.48% on CTA40, superior to both Mean Teacher (MT) and Mutual Consistency Network+ (MC-Net+). Ablation studies confirmed the efficacy of each module. Overall, the framework effectively leverages unlabeled volumetric data to yield precise vessel boundary delineations. Full article

Multi-Level Governance (MLG) frameworks effectively diagnose the complexity of regional renewable energy transitions but lack operational mechanisms for institutional implementation. This study develops a hybrid evidence-based architecture that integrates the Viable System Model (VSM) with computational intelligence and objective multi-criteria decision analysis. Methodologically, a systematic review following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol of 339 peer-reviewed articles (2015–2025) feeds a Latent Dirichlet Allocation (LDA) model that extracts K = 30 strategic topics (semantic coherence $C_v$ optimized over K = 5–40), operationalizing System 4 environmental sensing. In parallel, a 1 km² pixel-based spatial model integrates a National Conflict Index (INC, 2019–2024) with technical feasibility layers (Global Wind Atlas v4.0, Solargis, Servicio Geológico Colombiano) and applies CRITIC (CRiteria Importance Through Intercriteria Correlation) objective weighting and TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) prioritization (System 3 control). Robustness is confirmed through $\pm 10$–$20\%$ weight perturbation (Spearman > 0.92). Empirically, the framework reveals a localization paradox: approximately 68% of optimal wind zones (>9 m/s at 100 m hub height) in the Colombian Caribbean overlap with the highest national conflict quartile, narrowing a theoretical capacity exceeding 100 GW (50 GW offshore wind, 30 GW onshore wind, 42 GW solar PV, 1.17 GW geothermal) to roughly 24 GW of governance-viable capacity. Scenario calibration (Accelerated 80/20, Balanced 50/50, Justice-Oriented 30/70 technical/conflict weighting) demonstrates that System 5 normative orientation materially reshapes territorial prioritization. The framework advances VSM from a qualitative diagnostic metaphor to a reproducible governance architecture for high-variety regional contexts. Full article

Accurate retinal vessel segmentation supports quantitative vascular analysis in assessing ocular and systemic diseases. Yet, its clinical scalability is constrained by limited pixel-level annotations and domain shift across heterogeneous fundus datasets. Thus, this study proposes a domain-adaptive retinal vessel segmentation model (DA-VesselNet), a weakly supervised approach that transfers vessel-segmentation knowledge from annotated source datasets to the unannotated Retinal Fundus Multi-Disease Image Dataset (RFMiD). The model was trained on DRIVE, CHASE_DB1, and FIVES, and adapted to RFMiD. The ResNet50 encoder with an attention-gated U-Net decoder, confidence-aware pseudo-label supervision, and perturbation consistency regularisation were used for the training process. Results on the held-out CHASE_DB1 indicated that DA-VesselNet achieved a Dice score of 0.5778, an Intersection over Union (IoU) of 0.4082, and an Area Under the Curve (AUC) of 0.9518. On 200 held-out FIVES test images with different pathological features, it achieved a Dice score of 0.7447 and an AUC of 0.9733, outperforming the source-only baseline model. To assess adaptation independently of source-adjacent data, the model was further tested on STARE and HRF, two domains excluded entirely from source training. Adaptation improved Dice by 0.0469 on STARE and 0.0056 on HRF with AUC gains of 0.0140 and 0.0121, respectively. Ablation analysis identified source-anchored supervision as the dominant contributor to performance. The adapted model was subsequently applied to generate vessel pseudo-labels for the RFMiD target domain, providing a structural resource for future vessel-informed analysis. These findings demonstrate that DA-VesselNet offers a scalable solution for creating clinically relevant pseudo-labelled vessels in fundus imaging with limited annotations. Full article

Metallographic image segmentation underpins automated metallographic analysis, yet pixel-level annotation is costly. Metallographic microstructures exhibit dramatic size differences, including cross-scale structures such as large-scale matrix phases, grain boundary cementite networks, and acicular Widmanstätten structures. Existing feature extraction modules cannot balance the semantic integrity of large regions and the fine details of microstructures, causing missed small microstructures and blurred segmentation boundaries. Current semi-supervised methods only impose consistency constraints on perturbed input images at the final prediction output, ignoring multi-scale semantic features from decoder upsampling stages. This leads to noisy supervision signals, low-quality pseudo-labels, and poor generalization. To address these issues, this paper proposes a semi-supervised metallographic image segmentation model, EISC, integrating MT cross-network scale feature extraction and decoder multi-scale consistency constraints. It adaptively fuses multi-receptive-field features and introduces a regularization term to maintain the semantic consistency of multi-scale decoder outputs, improving pseudo-label quality. Comparative and ablation experiments verify that EISC effectively enhances the segmentation accuracy and robustness of metallographic images. Full article

Digital images, as a widely used multimedia carrier, are vulnerable to various security risks in open network environments. To enhance the security of digital images during storage and transmission, this paper proposes an image encryption algorithm based on a 2D folded sine-logistic map. First, the user master key and the SHA-256 hash of the plaintext image are combined to generate plaintext-related chaotic sequences. Second, a concentric-ring permutation is performed through three-ring partition, cyclic shifting, and route regrouping, with an additional channel crossover for RGB images to enhance inter-channel coupling. Finally, an enhanced hybrid diffusion mechanism with 4-bit nibble-swap perturbation is applied to strengthen pixel-value diffusion and local nonlinearity. Experimental results show that the average ciphertext entropy reaches 7.999, the adjacent-pixel correlation coefficient is reduced to 0.000016, and the NPCR/UACI values are close to their theoretical ideal levels. Combined with comprehensive security analyses, these findings demonstrate that the proposed scheme suppresses statistical characteristics and pixel correlations, resists brute-force, statistical, and differential attacks, and provides an efficient and secure solution for grayscale and color image transmission. Full article

Segmenting key components of space targets using Inverse Synthetic Aperture Radar (ISAR) images is an important interpretation task in space situational awareness. However, the scarcity of pixel-level annotated data, inter-class confusion caused by morphological differences among multiple target classes, and the absence of structural priors for components restrict the performance improvement in existing deep models on this task. Therefore, this paper proposes a Semi-Supervised Structural Prior-Guided Network (SSPNet). First, a Gated Manifold-Constrained Hyper-Connections Vision Transformer (GMHC-ViT) encoder is proposed to broaden the feature representation space via parallel multi-feature streams with adaptive gating, thereby alleviating inter-class confusion and enhancing cross-category generalization. Second, a Prior-Guided Module (PGM) is proposed to extract shape and edge priors of components, and it adaptively enhances the weakly activated channels of encoder features through cross-attention, thereby injecting structural knowledge independent of image quality into the segmentation process. Furthermore, to effectively leverage large amounts of unlabeled data, a strong perturbation strategy tailored to the characteristics of ISAR images is designed for consistency regularization. Experimental results on a simulated ISAR dataset containing 38 classes of space targets demonstrate that SSPNet outperforms existing methods and exhibits strong segmentation capability even under low signal-to-noise ratio (SNR) conditions. Full article

While existing AI-based microplastic monitoring studies predominantly rely on task-specific fine-tuned models, this study evaluates whether general-purpose multimodal foundation models with no spectroscopic instrumentation, minimal fine-tuning, and minimal computational resources can serve as accessible, low-cost tools for polymer classification under ecologically realistic field conditions with samples of plastic debris collected from coastal Sousse, Tunisia, a Mediterranean region experiencing anthropogenic pollution pressures. A curated dataset of 1080 high-resolution images was developed, representing six polymer groups (HDPE, LDPE, PA, PET, PP, PS, and mixed plastics). Fragments were imaged under standardized lighting conditions against natural sand backgrounds to preserve environmental realism. Each image was manually annotated using polygon-based boundaries to generate pixel-level segmentation masks and associated class labels, providing expert-validated ground truth for quantitative evaluation. Multimodal LLMs were evaluated using a composite scoring framework. Spatial accuracy was assessed using mean Intersection-over-Union (mIoU) against expert annotations, while polymer classification performance was measured using macro-averaged F1 scores across all categories. Model reliability was further evaluated through prompt stability testing and robustness analyses under controlled environmental perturbations designed to assess consistency across varying coastal imaging conditions. Results indicate that multimodal foundation models can distinguish plastic fragments from sand backgrounds, although performance varied across polymer classes and environmental perturbations. The weighted composite framework provides a structured approach for comparing model performance according to ecological monitoring objectives rather than computational metrics. These findings contribute to understanding the potential utility and current limitations of AI-based approaches for marine microplastic analysis and provide insights into coastal pollution patterns and ecosystem health within a One Health monitoring context. Full article

Background/Objectives: Robust brain-tumor segmentation on contrast-enhanced MRI remains limited by scanner-dependent intensity shifts, scarce annotations, and evaluation protocols that may leak patient-specific information. We propose BA-SwinMamba, a region-wise Bézier intensity augmentation framework built on Swin-UMamba, a selective state-space U-Net that combines hierarchical Swin-style visual modeling with Mamba’s linear-complexity long-range sequence representation. Materials and Methods: During training, independent monotonic or non-monotonic Bézier transfer functions are sampled for tumor and background regions, perturbing lesion-to-background contrast while preserving the binary mask geometry. Fourteen convolutional, transformer-based, and state-space segmentation models were evaluated on the Cheng brain-tumor dataset, comprising 3064 contrast-enhanced T1-weighted slices from 233 patients, using a strictly patient-level five-fold protocol. Single-source domain generalization was assessed by training only on Cheng and testing, without fine-tuning, on two independent target datasets. Results: BA-SwinMamba achieved 89.6% Dice, 82.0% IoU, and 5.9-pixel HD95 on the source domain, outperforming the plain Swin-UMamba backbone by 1.7 Dice points. The benefit was larger under domain shift: mean target-domain Dice increased from 72.7% with Swin-UMamba to 78.3% with BA-SwinMamba. Ablation analysis showed that replacing global Bézier augmentation with the proposed region-wise formulation added 1.5 Dice points. Conclusions: The method introduces no inference-time cost because augmentation is disabled after training, without modifying the deployed network or requiring target-domain labels during model optimization or tuning. The results indicate that lesion-aware intensity perturbation can improve cross-dataset robustness of Mamba-based 2D brain-tumor segmentation, while wider volumetric and multi-institutional validation remains necessary. Full article

Convolutional neural networks (CNNs) are widely used in medical image classification, yet their robustness to localized perturbations remains limited. This study evaluates one-pixel attacks on VGG16, MobileNetV2, and EfficientNetV2-B0 using brain tumor MRI images resized to 96 × 96 pixels. Each pixel was systematically perturbed across grayscale intensities, and model responses were analyzed in terms of vulnerability, recoverability, and pixel-level sensitivity. The relationship between prediction confidence and influential pixel locations was also examined. Results show that all models remain vulnerable to one-pixel perturbations despite high accuracy. Misclassified samples exhibit more successful attack locations, while correctly classified samples are more robust. Higher-intensity perturbations more often restore correct predictions in misclassified cases. A monotonic relationship is observed between prediction confidence and pixel sensitivity, where lower confidence corresponds to more influential pixels. Recovery points show spatially concentrated patterns. Overall, pixel-level sensitivity is more strongly associated with prediction correctness and local perturbations than with confidence. These findings are consistent across architectures and suggest that one-pixel analysis is useful for assessing CNN robustness in medical imaging. Full article

At 1 km resolution, NDVI projections for heterogeneous landscapes can appear spatially coherent in aggregate while concealing substantial class-level prediction weaknesses, a limitation that has received limited systematic attention in the NDVI projection literature. This study applies a four-component assessment workflow to Northeast China (NEC) for 2040 under SSP1-2.6, SSP2-4.5, and SSP5-8.5, integrating multi-stage model selection, land-cover-stratified validation, quantile-regression-based uncertainty characterization, and validation-priority ranking. Among three candidate tree-based models evaluated using spatial block cross-validation, temporal holdout validation, long-jump extrapolation, and climatic perturbation tests, LightGBM showed the most balanced and consistent performance, with spatial CV R² = 0.654 ± 0.123, temporal holdout R² = 0.710, and long-jump R² = 0.671, and was therefore selected for the 2040 projection. Projected regional mean NDVI increased modestly from 0.393 in 2020 to 0.414–0.417 across scenarios, with limited divergence among SSP pathways at this near-term horizon. Class-stratified validation of the 2020 holdout prediction revealed that global model performance masked strong class-level heterogeneity, with R² values ranging from 0.576 for Construction land to −0.886 for Unused land. Water bodies and Unused land exhibited negative R² values, indicating weak class-level predictive support relative to a simple class-mean benchmark. Residual decomposition showed that Water bodies combined high random error with elevated systematic deviation, whereas Unused land was mainly characterized by systematic bias, suggesting different needs for class-specific model improvement. The Uncertainty Risk Index (URI), derived from 95% prediction intervals, was highest in Construction land and lowest in Cropland across all scenarios. Integrating historical residuals with future URI-identified Water bodies, Unused land, and Construction land as the highest-priority classes for future targeted validation. These priorities arise from both limited class representation and intrinsic NDVI-related complexity, including low vegetation signal, mixed-pixel effects, and heterogeneous land-surface composition. These results demonstrate that land-cover-stratified error decomposition and uncertainty-informed priority ranking reveal class-specific projection limitations that aggregate accuracy metrics can conceal. Full article

Manipulated satellite imagery threatens analytic workflows, policy decisions, and trust in geospatial intelligence. Operational systems increasingly benefit from capabilities for both manipulation detection and manipulation-family attribution to support verification, triage, and downstream analysis. We present a unified benchmark for characterizing three representative manipulation families in geospatial imagery—generative manipulations, pixel-level perturbations, and adversarial patches—using a controlled, class-balanced design and 20 modern vision architectures spanning conventional, Earth-observation-pretrained, and vision-language models. Across architectures, the dominant failure boundary is between authentic imagery and subtle pixel-level perturbations, whereas generative manipulations and adversarial patches are generally more separable under matched in-domain conditions. Additional analyses reveal important generalization limitations under unseen manipulation variants and external-domain transfer, demonstrating that strong benchmark performance does not necessarily translate to reliable operational screening. The framework also enables systematic comparison of unified multi-attack and specialized detection strategies, providing insight into their relative strengths and limitations. Rather than proposing a new defense, this work provides a reproducible methodology for characterizing manipulation artifacts, model failure modes, and deployment-relevant screening behavior in geospatial imagery, with applications to analyst triage, verification workflows, and trustworthy use of satellite data. Full article

The growing need for secure image transmission across public networks requires robust encryption algorithms. Traditional chaos-based image ciphers typically have a small key space, weak avalanche behavior, or are susceptible to differential cryptanalysis. To overcome such inadequacies, this paper suggests a new adaptive image cryptosystem that combines a fractional-order memristive chaotic engine and a non-linear hybrid encryption kernel. The system uses piano-inspired feedback; the keystream generator dynamically adapts to the previously encrypted pixel, enabling powerful Cipher Block Chaining (CBC)-style chaining and content-dependent diffusion. A four-dimensional memristive system is solved by the use of fractional-order calculus, which gives an ultra-large key space (>10⁸⁰) and very high sensitivity to initial conditions—confirmed by a positive largest Lyapunov exponent (1.7199). The encryption kernel maps the traditional Exclusive OR (XOR) with the reversible two-step operation: the modular addition of the plaintext with the first keystream byte and the XOR with the second keystream one, both of which increase non-linearity and confusion. Large-scale experiments with six standard 256 × 256 colour images indicate almost ideal entropy (7.9994), Number of Pixel Change Rate (NPCR) which is 99.62, Unified Average Changing Intensity (UACI) which is 33.43, correlation coefficients are near to zero, very low Gray-Level Co-occurrence Matrix (GLCM) homogeneity (≈0.017) and high contrast (≈4843) and low energy (≈0.006 The ciphertext passes seven National Institute of Standards and Technology (NIST) SP-800-22 statistical tests, is extremely sensitive to keys (a perturbation of 1 × 10⁻¹⁴ alters >99.6% of ciphertext) and resists chosen-plaintext and known-plaintext attacks. Decryption has linear time complexity O(N), and average encryption and decryption times are 3.40 s and 2.75 s for 256 × 256 images. The proposed cryptosystem provides an attractive security–performance trade-off that can be used in high-security systems like medical image protection, privacy-preserving multimedia transmission, and secure cloud storage. Full article