Search Results (200)

Low-light image enhancement (LLIE) remains a challenging task due to the scarcity of paired training data and the complex signal-dependent noise inherent in low-light scenes. To address these issues, this paper proposes a fully unsupervised framework named Wavelet-Guided Zero-Reference Diffusion (WZD) for natural low-light image restoration. WZD leverages an ImageNet-pre-trained diffusion prior and a multi-scale representation of the Discrete Wavelet Transform (DWT) to restore natural illumination from a single dark image. Specifically, the input low-light image is first processed by a Practical Exposure Corrector (PEC) to provide an initial robust luminance baseline. It is then converted from the RGB to the YCbCr color space. The Y channels of the input image and the current diffusion estimate are decomposed into four orthogonal sub-bands—LL, LH, HL, and HH—and fused via learnable, step-wise weights while preserving structural integrity. An exposure control loss and a detail consistency loss are jointly employed to suppress over/under-exposure and preserve high-frequency details. Unlike recent approaches that rely on complex supervised training or lack physical guidance, our method integrates wavelet guidance with a zero-reference learning framework, incorporates the PEC module as a physical prior, and achieves significant improvements in detail preservation and noise suppression without requiring paired training data. Comprehensive experiments on the LOL-v1, LOL-v2, and LSRW datasets demonstrate that WZD achieves a superior or competitive performance, surpassing all referenced unsupervised methods. Ablation studies confirm the critical roles of the PEC prior, YCbCr conversion, wavelet-guided fusion, and the joint loss function. WZD also enhances the performance of downstream tasks, verifying its practical value. Full article

Haze significantly reduces visibility in critical applications such as autonomous driving, surveillance, and firefighting, making its removal essential for safety and reliability. Motivated by the limited robustness of the existing methods under non-uniform haze conditions, this study introduces a novel regression-based dehazing model that simultaneously incorporates the atmospheric light constant, transmission map, and scattering coefficient for improved restoration. Instead of relying on complex deep networks, the model leverages brightness–saturation cues and regression-driven scattering estimation with localized haze detection to reconstruct clearer images efficiently. Evaluated on the RESIDE dataset, the approach consistently surpasses state-of-the-art techniques including Dark Channel Prior, AOD-Net, FFA-Net, and Single U-Net, achieving SSIM = 0.99, PSNR = 22.25 dB, VIF = 1.08, and the lowest processing time of 0.038 s, demonstrating both accuracy and practicality for real-world deployment. Full article

Cytokinins (CKs) are central regulators of leaf senescence, yet their cultivar-specific functions in cereals remain insufficiently understood. Here, we examined dark-induced senescence (DIS) in three barley (Hordeum vulgare L.) cultivars: Carina, Lomerit, and Bursztyn, focusing on responses to exogenous benzyladenine (BA) and inhibition of endogenous CK biosynthesis via the mevalonate (MVA) pathway using lovastatin (LOV). Bursztyn, a winter cultivar, displayed a previously uncharacterized stay-green phenotype, characterized by delayed chlorophyll and protein degradation and reduced sensitivity to BA with respect to chlorophyll retention. In contrast, Carina (spring) senesced rapidly but exhibited strong responsiveness to BA. Lomerit (winter) showed an intermediate phenotype, combining moderate natural resistance to senescence with clear responsiveness to BA. CK application suppressed SAG12 cysteine protease accumulation in all cultivars, serving as a marker of senescence and N remobilization, stabilized photosystem II efficiency, preserved photosynthetic proteins, and alleviated oxidative stress without promoting excessive energy dissipation. Although BA only partially mitigated the decline in net CO₂ assimilation, it sustained ribulose-1,5-bisphosphate regeneration, supported electron transport, and stabilized Rubisco and Rubisco activase. Moreover, LOV-based inhibition of the MVA pathway of CK biosynthesis revealed that endogenous CK contributions to senescence delay were most pronounced in Lomerit, moderate in Bursztyn, and negligible in Carina, indicating genotype-specific reliance on MVA-versus methylerythritol phosphate (MEP) pathway-derived CK pools. Collectively, these findings identify Bursztyn as a novel genetic resource for stay-green traits and demonstrate that BA delays DIS primarily by maintaining photosynthetic integrity and redox balance. The results highlight distinct regulatory networks shaping CK-mediated senescence responses in cereals, with implications for improving stress resilience and yield stability. Full article

Artificial intelligence (AI) is an important tool for improving agricultural tasks. In particular, object detection methods based on convolutional neural networks (CNNs) enable the detection and classification of objects directly in the field. Combined with unmanned aerial vehicles (UAVs, drones), these methods allow efficient crop monitoring. The primary challenge is to develop models that are both accurate and feasible under real-world conditions. This study addresses this challenge by evaluating marigold flower detection using three groups of CNN detectors: canonical models, including YOLOv2, Faster R-CNN, and SSD with their original backbones; modified versions of these detectors using DarkNet-53; and modern architectures, including YOLOv11, YOLOv12, and the RT-DETR. The dataset consisted of 392 images from marigold fields, which were manually labeled and augmented to a total of 940 images. The results showed that YOLOv2 with DarkNet-53 achieved the best performance, with 98.8% mean average precision (mAP) and 97.9% F1-score (F1). SSD and Faster R-CNN also improved, reaching 63.1% and 52.8%, respectively. Modern models obtained strong results: YOLOv11 and YOLOv12 reached 96–97%, and RT-DETR 93.5%. The modification of YOLOv2 allowed this classical detector to compete directly with, and even surpass, recent models. Precision–recall (PR) curves, F1-scores, and complexity analysis confirmed the trade-offs between accuracy and efficiency. These findings demonstrate that while modern detectors are efficient baselines, classical models with updated backbones can still deliver state-of-the-art results for UAV-based crop monitoring. Full article

Artificial light has been shown to enhance the fishing efficiency of fishing gear by attracting marine organisms. This study introduces a novel approach by incorporating noctilucent materials into crab pots and evaluates their effects on catch performance. Based on the crab pots commonly used on the coast, four types of crab pots were tested: ordinary crab pots (Con-pot), ordinary crab pots equipped with noctilucent sticks (Exp-pot 1), crab pots equipped with noctilucent nets (Exp-pot 2), and crab pots equipped with both noctilucent nets and sticks (Exp-pot 3). The results showed that the noctilucent material exhibits 6 h persistent emission in darkness after just 10 min of solar charging. Exp-pot 3 could significantly enhance fishing efficiency, which increased by 63.84% compared to the Con-pot. The proportion of crabs in Exp-pot 3 was the highest (86.35%), and the individual weight of crabs in Exp-pot 3 was the heaviest (61.5 g), which was 38.30% heavier than that in the Con-pot. Notably, Exp-pots 2 and 3 demonstrated superior selectivity with higher W₅₀ values (53.01 g and 54.49 g), narrower SRs (33.04–72.98 g and 32.95–76.03 g), effectively balancing target catch retention with undersized crab release, indicated that noctilucent nets exhibited stronger weight selectivity for crabs compared to noctilucent sticks. These results demonstrate that functional materials have broad potential applications in fishing gear, which could enhance the catch efficiency and individual size of crab caught. Full article

Land salinization severely limits the development of agriculture, and the growing global population poses a serious challenge to food security. As an abiotic stress factor limiting photosynthesis in potatoes (Solanum tuberosum L.), alkaline salt stress significantly impacts their photosynthetic activity. In this study, potted seedlings of the ‘Atlantic’ variety were planted in the pots. Sodium bicarbonate (NaHCO₃) was incorporated into the dry soil within the pots at four distinct concentration levels: 0 mmol/L, 20 mmol/L, 40 mmol/L, and 60 mmol/L. The findings indicated that at a concentration of 60 mmol/L, the initial fluorescence (Fo) exhibited its peak value. At this concentration, NaHCO₃ stress induced a significant decline in several parameters: variable fluorescence (Fv), the chlorophyll fluorescence ratio (Fv/Fm), dark-adapted maximum fluorescence (Fm), the Fv/Fo ratio, and overall plant performance. Compared to the control CK, the values of Fv, Fv/Fm, Fm, and Fv/Fo decreased by 42.36%, 20.44%, 54.1%, and 61.97%, respectively. At a stress concentration of 60 mmol/L, NaHCO₃ stress exhibited a more pronounced inhibition of chlorophyll synthesis. Under T3 treatment at this stress concentration, the contents of chlorophyll a, chlorophyll b, and total chlorophyll a/b were significantly lower than the control group (CK), decreasing by 46.29%, 54.3%, and 48.56%, respectively. The T2 treatment showed the next most pronounced reduction, with levels 33.26%, 45.75%, and 36.79% lower than CK, respectively. After a brief increase in the intercellular CO₂ concentration (Ci) in photosynthetic gas exchange, the net photosynthetic rate (Pn), stomatal conductance (Gs), and transpiration rate (Tr) decreased significantly with the gradual increase in concentration and prolongation of time. The expression levels of genes related to some subunits of photosystem II and photosystem I were down-regulated under stress, while the expressions of genes related to Fd and FNR were also down-regulated to varying degrees. In this study, photosynthetic activities such as fluorescence parameters, chlorophyll content, and photosynthetic gas exchange were measured, along with 16 key photosynthetic genes of potato plants. The aim was to explore the effects of alkaline salt stress on potato photosynthesis and its related mechanisms. The research outcomes contribute to a better understanding of potato’s adaptive responses to alkaline stress, potentially informing future efforts in crop improvement and saline agriculture management. Full article

Dark matter dominates the matter content of the Universe, yet its particle nature remains elusive. Among the promising multi-messenger astronomy dark matter candidates are weakly interacting massive particles and superheavy dark matter, both of which may manifest themselves in cosmic ray, $\gamma$-ray, and neutrino signatures through annihilation or decay. Here, we explore potential multi-messenger signals from these candidates in galaxy clusters of the Shapley Supercluster—one of the most massive known structures in the local Universe (located at a distance of ∼200 Mpc and containing over ${10}^{16}{M}_{\odot }$ of dark matter). Using the CLUMPY code, we model $\gamma$-ray and neutrino fluxes for weakly interacting massive particle masses between 0.1 and 100 TeV across various final states, comparing the predictions with the sensitivities of current and forthcoming observatories, including CTAO, IceCube, and KM3NeT. For superheavy dark matter scenarios with masses from ${10}^{19}$ to ${10}^{28}$ eV, we employ HDMSpectra code to compute ultra-high-energy cosmic ray proton and neutrino fluxes in the ranges available for observations using present (Pierre Auger Observatory, IceCube, KM3NeT) and future (GRAND, GCOS, etc.) instruments. Full article

We investigate the impact of dataset augmentation and synthetic generation techniques on the accuracy of supervised audio classification based on state-of-the-art neural networks used as classifiers. Dataset augmentation techniques are applied upon the raw sound and its transformed image format. Specifically, sound augmentation techniques are applied prior to spectral-based transformation and include time stretching, pitch shifting, noise addition, volume controlling, and time shifting. Image augmentation techniques are applied after the transformation of the sound into a scalogram, involving scaling, shearing, rotation, and translation. Synthetic sound generation is based on the AudioGen generative model, triggered through a series of customized prompts. Augmentation and synthetic generation are applied to three sound categories: (a) human sounds, (b) animal sounds, and (c) sounds of things, with each category containing ten sound classes with 20 samples retrieved from the ESC-50 dataset. Sound- and image-orientated neural network classifiers have been used to classify the augmented datasets and their synthetic additions. VGGish and YAMNet (sound classifiers) employ spectrograms, while ResNet50 and DarkNet53 (image classifiers) employ scalograms. The streamlined AI-based process of augmentation and synthetic generation, enhanced classifier fine-tuning and inference allowed for a consistent, multicriteria-comparison of the impact. Classification accuracy has increased for all augmentation and synthetic generation scenarios; however, the increase has not been uniform among the techniques, the sound types, and the percentage of the training set population increase. The average increase in classification accuracy ranged from 2.05% for ResNet50 to 9.05% for VGGish. Our findings reinforce the benefit of audio augmentation and synthetic generation, providing guidelines to avoid accuracy degradation due to overuse and distortion of key audio features. Full article

Bats are ecologically vital mammals, serving as pollinators, seed dispersers, and bioindicators of ecosystem health. Many species inhabit natural caves, which offer optimal conditions for survival but present challenges for direct ecological monitoring due to their dark, complex, and inaccessible environments. Traditional monitoring methods, such as mist-netting, are invasive and limited in scope, highlighting the need for non-intrusive alternatives. In this work, we present a portable multisensor platform designed to operate in underground habitats. The system captures multimodal data, including near-infrared (NIR) imagery, ultrasonic audio, 3D structural data, and RGB video. Focusing on NIR imagery, we evaluate the effectiveness of the YOLO object detection framework for automated bat detection and counting. Experiments were conducted using a dataset of NIR images collected in natural shelters. Three YOLO variants (v10, v11, and v12) were trained and tested on this dataset. The models achieved high detection accuracy, with YOLO v12m reaching a mean average precision (mAP) of 0.981. These results demonstrate that combining NIR imaging with deep learning enables accurate and non-invasive monitoring of bats in challenging environments. The proposed approach offers a scalable tool for ecological research and conservation, supporting population assessment and behavioral studies without disturbing bat colonies. Full article

Hydrogen is key to achieving a net-zero carbon future, yet current production remains predominantly fossil-based. Biohydrogen derived from agricultural residues represents a sustainable alternative aligned with circular economy principles. While several studies have assessed the bioenergy potential from agricultural residues in various African countries, their potential in Togo remains largely unexplored. This study employed an exploratory mixed-methods approach to quantify residue availability, evaluate production pathways, and estimate potential biohydrogen yields. Secondary data on crop production from the Food and Agriculture Organization (FAO) and theoretical conversion factors were used to assess the availability of agricultural residues from the eight major crops in Togo, resulting in a residue potential of 7.95 million tons per year. Considering ecological and competing aspects of residue utilization, a sustainable share of 3.1 to 6.6 million tons was estimated to be available for biohydrogen production, depending on the residue recoverability assumptions. A multi-criteria decision analysis (MCDA) was used to evaluate different biohydrogen production processes, identifying dark fermentation as the most suitable due to its low energy requirements and decentralized applicability. The theoretical biohydrogen potential was estimated at 20,991–42,293 tons per year (2.5–5.1 PJ per year) based on biochemical residue composition data and stoichiometric calculations. This study established a baseline assessment of biohydrogen potential from agricultural residues in Togo, offering a methodological framework for assessing biohydrogen potential in other regions. The results also underscore the need for site-specific data to reduce uncertainty and support evidence-based energy planning. Full article

This paper proposes a novel Denoiser and Low-Frequency Enhancer Network (DLFE-Net) for Low-Light Image Enhancement (LLIE). The DLFE-Net addresses two key challenges: (1) overexposure and detail loss in local areas during enhancement, and (2) the effective removal of inherent noise in low-light images. Specifically, the input RGB image is first converted to the HVI color space. The intensity (I) and color (H, V) maps are then enhanced and denoised separately, i.e., preserving details and removing noise. For preserving details, the Low-Frequency Illumination Enhancer (LFIE) module isolates and processes the image’s low-frequency information. This targeted approach effectively mitigates local overexposure and preserves fine details during enhancement. For removing noise, the Multi-Scale Gated Denoiser (MSGD) module performs denoising through strong preservation after predicting image noise. Comprehensive experiments were conducted on three benchmark datasets (LOL, SICE, Sony-Total-Dark) and five unpaired datasets. Both qualitative and quantitative analyses demonstrated the superiority of DLFE-Net over state-of-the-art methods. Moreover, ablation studies demonstrated the effectiveness of each module in DLFE-Net. Full article

Underwater fish image segmentation is the key technology to realizing intelligent fisheries and ecological monitoring. However, the problems of light attenuation, blurred boundaries, and low contrast caused by complex underwater environments seriously restrict the segmentation accuracy. In this paper, RetinalConet, an underwater fish segmentation network based on bionic retina dual-channel and multi-module cooperation, is proposed. Firstly, the bionic retina dual-channel module is embedded in the encoder to simulate the separation and processing mechanism of light and dark signals by biological vision systems and enhance the feature extraction ability of fuzzy target contours and translucent tissues. Secondly, the dynamic prompt module is introduced, and the response of key features is enhanced by inputting adaptive prompt templates to suppress the noise interference of water bodies. Finally, the edge prior guidance mechanism is integrated into the decoder, and low-contrast boundary features are dynamically enhanced by conditional normalization. The experimental results show that RetinalCoNet is superior to other mainstream segmentation models in the key indicators of mDice, reaching 82.3%, and mIou, reaching 89.2%, and it is outstanding in boundary segmentation in many different scenes. This study achieves accurate fish segmentation in complex underwater environments and contributes to underwater ecological monitoring. Full article

Downy mildew, caused by the oomycete Pseudoperonospora cubensis, is the most destructive foliar disease of cucumbers. While partially resistant slicer cultivars (with spined fruits) are commercially available, no resistant Beit Alpha cultivars (characterized by smooth, dark green fruit) have been developed to date. Here, we report the successful breeding of downy mildew-resistant Beit Alpha cucumber lines. Resistance was transferred from the wild Sikkim cucumber accessions PI 197088 and PI 330628 (characterized by round fruit, with heavily netted brown rind). The resistance and fruit phenotype were restored through backcrosses to elite commercial susceptible cultivars. Due to the recessive nature of the resistance genes and their distribution across multiple chromosomes, the breeding program required multiple backcrosses and stringent selections for both resistance and fruit type. Full article

Organic photodetectors (OPDs) offer considerable promise for low-power, solution-processable biosensing and imaging applications; however, their performance remains limited by spectral mismatch and interfacial trap states. In this study, a highly sensitive polymer photodiode was developed via trace incorporation (0.8 wt%) of InP/ZnSe/ZnS quantum dots (QDs) into a PTB7-Th:PC₇₁BM bulk heterojunction (BHJ) matrix. This QD doping approach enhanced the external quantum efficiency (EQE) across the 540–660 nm range and suppressed the dark current density at −2 V by passivating interface trap states. Despite a slight decrease in optical absorption at the optimized composition, the internal quantum efficiency (IQE) increased significantly from ~80% to nearly 95% resulting in a net EQE improvement. This suggests that QD incorporation improved charge transport without compromising charge separation efficiency. As a result, the device achieved a specific detectivity (D*) of 1.8 × 10¹³ Jones, representing a 93% improvement over binary BHJs, along with an ultra-low dark current density of 7.76 × 10⁻¹⁰ A/cm². Excessive QD loading, however, led to optical losses and increased dark current, underscoring the need for precise compositional control. Furthermore, the enhanced detectivity led to a 4 dB improvement in the signal-to-noise ratio (SNR) of photoplethysmography (PPG) signals in the target wavelength range, enabling more reliable biophotonic sensing without increased power consumption. This work demonstrates that QD-based spectral and interfacial engineering offers an effective and scalable route for advancing the performance of OPDs, with broad applicability to low-power biosensors and high-resolution polymer–QD imaging systems. Full article

Simple Baseline has become a dominant benchmark in human pose estimation (HPE) due to its excellent performance and simple design. However, its “strong encoder + simple decoder” architectural paradigm suffers from two core limitations: (1) its non-branching, linear deconvolutional path prevents it from leveraging the rich, fine-grained features generated by the encoder at multiple scales and (2) the model lacks explicit prior knowledge of both the absolute positions and structural layout of human keypoints. To address these issues, this paper introduces AFJ-PoseNet, a new architecture that deeply enhances the Simple Baseline framework. First, we restructure Simple Baseline’s original linear decoder into a U-Net-like multi-scale fusion path, introducing intermediate features from the encoder via skip connections. For efficient fusion, we design a novel Attention Fusion Module (AFM), which dynamically gates the flow of incoming detailed features through a context-aware spatial attention mechanism. Second, we propose the Joint-Aware Positional Encoding (JAPE) module, which innovatively combines a fixed global coordinate system with learnable, joint-specific spatial priors. This design injects both absolute position awareness and statistical priors of the human body structure. Our ablation studies on the MPII dataset validate the effectiveness of each proposed enhancement, with our full model achieving a mean PCKh of 88.915, a 0.341 percentage point improvement over our re-implemented baseline. On the more challenging COCO val2017 dataset, our ResNet-50-based AFJ-PoseNet achieves an Average Precision (AP) of 72.6%. While this involves a slight trade-off in Average Recall for higher precision, this result represents a significant 2.2 percentage point improvement over our re-implemented baseline (70.4%) and also outperforms other strong, publicly available models like DARK (72.4%) and SimCC (72.1%) under comparable settings, demonstrating the superiority and competitiveness of our proposed enhancements. Full article