Image deblurring aims to restore high-quality images from blurry ones. Although existing methods have achieved remarkable progress in restoration performance, most approaches still mainly rely on designing complex modules in the spatial domain, leading to the gradual saturation of performance improvements. This paper proposes a novel CNN–Transformer hybrid architecture for image deblurring, dubbed DSCTNet, by analyzing the differences between blurry and sharp images in the fourier and wavelet domains. Specifically, DSCTNet comprises two stages. The first stage is the local restoration stage, which progressively removes different levels of blur through multiple iterations. During this process, the multi-scale features generated at each iteration are fused into enhanced prior features, which are then forwarded along with the restored outputs to the second stage. The second stage is the global restoration stage, which employs a dual-branch design in the spatial and wavelet domains and incorporates a Transformer-based self-attention mechanism to capture long-range dependencies. By leveraging the prior features from the first stage, this stage effectively compensates for the Transformer’s limitations in local modeling, further improving structural restoration and detail reconstruction based on the preceding stage’s results. Additionally, a dual-domain feature enhancement module is designed to extract features at different frequencies in the fourier domain and features with multiple receptive fields in the spatial domain, subsequently fusing these features to enhance representational richness and diversity. Extensive experiments demonstrate that the proposed DSCTNet achieves superior image restoration performance compared with existing state-of-the-art methods.
Current automated mastitis detection methods typically identify dairy cows only in a fixed standing posture, which limits their scope and practical flexibility. To address potential occlusion issues across different postures, we divided the key anatomical regions into three areas (the eye, the back surface of the udder (BSU) and the lower surface of the udder (LSU)) and developed an infrared thermography (IRT)-based automated diagnostic system for cows in various postures within a lactation barn. First, a three-stage image enhancement method was applied to extract contour and texture features from the IRT images. Next, the You Only Look Once v8 Nano (YOLOv8n) model was improved by integrating Dynamic Snake Convolution to better capture subtle and complex texture patterns. We further optimised the weight distribution of contour and texture features using an efficient multi-scale attention module to reduce the loss of critical information in the deep network. Structural improvements included adding a P2 detection head to focus on contour features in the target regions. Finally, we built three machine learning models to diagnose mastitis using the maximum body temperatures of these critical regions. Results showed that the three-stage image enhancement effectively enriched IRT details, strengthened contour and texture features and improved detection confidence for the eye, BSU and LSU by 0.025, 0.05 and 0.04, respectively. The improved YOLOv8n model achieved top performance, with precision (P) of 94.2%, 97.8% and 96.1%; recall (R) of 96.6%, 94.1% and 89.7%; and average precision at an intersection-over-union of 50% (AP@0.5) of 94.3%, 93.7% and 94.2% for the eye, BSU and LSU, respectively. Compared with the baseline YOLOv8n model, the enhanced version improved P, R and AP@0.5 metrics by 2%∼5.4% across the three regions of interest. Among the diagnostic models, random forest achieved the highest accuracy at 92.31%. This method broadens the application of automated mastitis detection and provides a reference framework for building automatic monitoring systems for mastitis in feeder barns.
Soybean lipophilic protein (LP) is composed of beta-conglycinin (7S), glycinin (11S), and oil body proteinphosphatidylcholine (OBPs-PC), but the interactions among its components and their effects on foaming properties remain unclear. In this study, we reconstituted these components at the native LP ratio and systematically investigated the regulatory mechanism of protein-protein interactions on foaming properties. The results showed that the nature of component interactions determined foaming properties. 7S and 11S exhibited a synergistic effect: compared with the individual components, their combination underwent conformational rearrangement, characterized by decreased particle size, transition from beta-sheet to random coil, increased free sulfhydryl content, reduced surface hydrophobicity, and fluorescence quenching. These changes enhanced interfacial adsorption and foam stability, making 7S + 11S exhibit the optimal foaming ability (136.44%) and stability (58.42%). In contrast, OBPs-PC, due to its strong hydrophobicity and large-sized aggregates, inhibited interfacial adsorption, causing a significant decrease in the foaming ability of OBPs-PC-containing complexes. Molecular interaction analysis indicated that hydrogen bonds and hydrophobic interactions were the main non-covalent forces driving complex formation. Rheological results confirmed that 7S + 11S foam had the highest storage modulus and the widest linear viscoelastic region. In conclusion, 7S and 11S achieved optimal foaming properties through conformational rearrangement and synergistic interfacial behavior, while the strong hydrophobicity and largesized aggregates of OBPs-PC disrupted interfacial equilibrium and inhibited foaming. This study provides a theoretical basis for expanding the application of LP as efficient gas-water interface stabilizers and foaming agents.
Soil organic carbon (SOC) sustains ecosystem productivity, soil health, and sequesters atmospheric CO2. Straw return (StrawR) effectively compensates for carbon (C) losses by SOC mineralization in croplands. Quantifying the straw-derived SOC and straw conversion efficiency (SCE; the percentage of straw C converted to SOC) enables a direct assessment of C sequestration potential. This study integrates 13C isotopic tracer data with machine learning approaches to evaluate straw-derived SOC and SCE. A random forest model was further used to identify the key environmental and management drivers controlling straw-derived SOC and SCE, and to extrapolate their spatial patterns at the global scale. Straw-derived SOC content decreased over time, primarily due to the relative accumulation of recalcitrant compounds. Such dynamics are typically mediated by changes in microbial metabolic strategies in response to shifting resource availability. Random forest analysis identified StrawR amount, straw particle size, and soil bulk density (BD) as the key drivers of straw-derived SOC content (IncMSE percentages: 42%, 20%, and 19%, respectively). High soil BD potentially reduces soil aeration and suppresses microbial metabolic capacity, reducing C sequestration. Machine learning predictions indicate a straw C residual ratio of 17 ± 3.4% after 1 year and a global average SCE of 10 ± 1.1% after 5 years of StrawR, which supports our hypothesis that initial StrawR practices elevated C sequestration potential and SCE compared with prolonged StrawR application. Assuming 100% global adoption, StrawR offers a theoretical maximum biophysical potential of 1.7 Pg C yr−1 over five years. This maximum capacity would theoretically offset 52% of agricultural CO2 emissions and 16% of total anthropogenic CO2 emissions. This study addresses critical gaps in straw conversion dynamics and updated estimates of C sequestration capacity, highlighting the contribution of StrawR as a climate change mitigation strategy.
3D printing has emerged as a promising technology for developing nutritional foods to meet the needs of people with dysphagia. However, a single hydrogel or oleogel cannot meet the strict performance requirements for 3D printing, while bigels are ideal inks by combining the merits of both gels. In this study, soybean protein isolate (SPI)–rice bran oil (RBO) bigels were prepared at oleogel/hydrogel ratios of 30:70, 40:60, 50:50, 60:40, and 70:30. The samples were characterized by confocal laser scanning microscopy (CLSM), Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD), differential scanning calorimetry (DSC), rheology, texture profile analysis (TPA), low-field nuclear magnetic resonance (LF-NMR), liquid binding capacity (LBC), oxidative stability, International Dysphagia Diet Standardisation Initiative (IDDSI) tests, and manual extrusion using a 3D printing pen. Increasing oleogel content promoted phase evolution from oil-in-water (O/W) to bicontinuous and water-in-oil (W/O) structures, enhanced hydrogen-bonding-related spectral changes and crystallinity, and reduced free-water mobility. All formulations showed shear-thinning behavior and elastic-dominant networks. Among them, BG6 (60:40) showed the most balanced performance, including the highest thermal stability (Tpeak = 55.17 °C), high LBC (98.68%), relatively slow lipid oxidation, IDDSI Level 5 compliance, and visually stable F-, Z-, and cube-shaped structures under manual extrusion. These results suggest that SPI-RBO bigels, particularly BG6, have potential as plant-based edible inks for dysphagia-oriented 3D printed foods under the tested conditions.
