Bilinear Interpolation Research Articles

An error compensation method for on-machine measuring blade with industrial robot

To achieve the seamless integration of precision grinding blade with timely measured data, an on-machine measurement system employing industrial robot is developed. The limitation to its widespread application lies in the accuracy of measurement, prompting the proposal of an error compensation method to address the measurement error within the implemented blade on-machine measurement system. To address the pre-travel error in the measuring device of the on-machine measurement system, an analysis of the factors influencing the pre-travel error of the contact probe is conducted. Subsequently, pre-travel error data is gathered utilizing the experimental calibration method involving a standard ball, and a prediction model employing the bilinear interpolation algorithm is established to facilitate error compensation. Employing pre-travel error compensation for the analysis of the operating body error in the on-machine measurement system, a spherical center distance constraint error model is derived through closed-loop kinematic calibration of the six-degrees-of-freedom industrial robot. Kinematic parameter error identification is conducted using a hybrid algorithm combining the L-M algorithm and the adaptive factor double-variable DE algorithm. This approach diminishes the spherical center distance measurement error, reducing it from 0.081 mm to 0.016 mm. Subsequently, experiment is conducted to measure blade machining allowances using an on-machine robotic measurement system, comparing the obtained data with measurement from a blue light scanner and a coordinate measuring machine (CMM). The results reveal average absolute deviations of 0.020 mm, 0.015 mm, and 0.016 mm in the three blade cross sections for the on-machine robotic measurement system and the blue light scanner, respectively. Correspondingly, the average absolute deviations for the on-machine robotic measurement system and the CMM in the three blade sections are 0.028 mm, 0.029 mm, and 0.029 mm. Moreover, the on-machine measurement system demonstrates commendable measurement repeatability, with a standard deviation of 0.003 mm.

Evaluation of the PM2.5 concentrations in South America: Climatological patterns and trend analysis

This study assessed the concentration of particulate matter with an aerodynamic diameter of 2.5 (PM2.5) over South America (SA) based on climatological patterns and trend analysis. This study used monthly PM2.5 data from the Copernicus Atmosphere Monitoring Service (CAMS) at the European Centre for Medium-range Weather Forecasts from 2003 to 2022. The bilinear interpolation method was applied to the data resampling from 0.75° × 0.75°–0.25° × 0.25° spatial resolution, and subsequently, the climatological analysis was performed on seasonal and annual scales of PM2.5 concentrations for the 20-year series. The assessment comprised three main stages: 1) descriptive statistics, 2) seasonal and annual climatology patterns, and 3) trend analysis (Mann-Kendall - MK and Pettitt tests). Results indicated that the largest annual PM2.5 concentrations were found in the Amazon region and northern Paraguay due the El-Niño South Oscillation events (2003–2005, 2007, 2010, 2020, and 2022). The annual average PM2.5 concentrations exceeded the WHO recommended limit (5 μg m−3), ranging from 15.57 μg m−3 (1999) and 22.74 μg m−3 (2007 and 2010). The highest PM2.5 concentrations were observed in the Amazon and Chile (Santiago), reaching 160.4 μg m−3 annually and 177.8 μg m−3 seasonally (spring). Elevated seasonal PM2.5 concentrations were attributed to increased deforestation and wildfires associated with meteorological systems such as the Bolivian high and subtropical jets during dry periods (winter and spring). The MK test revealed a significant reduction in PM2.5 concentrations between 2003 and 2022, observed in Argentina (north), Brazil (arc of reforestation), Bolivia (south), and Chile (north). Reduction values varied between 0.5 and 10 μg m−3.year−1 annually and during winter, attributed to no-tillage practices in regions cultivating commodities. The Pettitt test also identified significant structural changes occurring in 2006–2008 and 2010–2011, aligning with areas highlighted in the MK test. This study provides valuable insights for air quality and environmental monitoring managers to mitigate PM2.5 pollution in SA.

Association of metabolic signatures of air pollution with MASLD: Observational and Mendelian randomization study

Background & AimsTo identify metabolic signatures associated with exposure to ambient air pollution and to explore their associations with risk of metabolic dysfunction-associated steatotic liver disease (MASLD). MethodsWe utilized data from the UK Biobank Cohort. Annual mean concentrations of PM2.5, PM10, NO2 and NOx were assessed for each participant using bilinear interpolation. The Elastic Net regression model was used to identify metabolites associated with four air pollutants and to construct metabolic signatures, respectively. Associations between air pollutants, metabolic signatures and MASLD were analyzed using Cox models. Mendelian randomization (MR) analysis was used to examine potential causality. Mediation analysis was employed to examine the role of metabolic signatures in the association between air pollutants and MASLD. ResultsA total of 244,842 participants from the UK Biobank were included in this analysis. We identified 87, 65, 76, and 71 metabolites as metabolic signatures of PM2.5, PM10, NO2, and NOx, respectively. Metabolic signatures were associated with risk of MASLD, with hazard ratios (HRs) and 95% confidence intervals (95% CIs) were 1.10 (1.06, 1.14), 1.06 (1.02, 1.10), 1.24 (1.20, 1.29) and 1.14 (1.10, 1.19). The four pollutants were associated with increased risk of MASLD, with HRs (95% CIs) of 1.03 (1.01, 1.05), 1.02 (1.01, 1.04), 1.01 (1.01, 1.02) and 1.01 (1.00, 1.01). MR analysis indicated an association between PM2.5, NO2 and NOx-related metabolic signatures and MASLD. Metabolic signatures mediated the association of PM2.5, PM10, NO2 and NOx with MASLD. ConclusionThere may be association between PM2.5, PM10, NO2 and NOx-related metabolic signatures and MASLD, and metabolic signatures mediate the increase of PM2.5, PM10, NO2 and NOx in the risk of MASLD. Impact and implicationsAir pollution is a significant public health issue and an important risk factor for metabolic dysfunction-associated steatotic liver disease (MASLD), however, the mechanism by which air pollution affects MASLD remains unclear. Our study used integrated serological metabolic data of 251 metabolites from a large-scale cohort study to demonstrate that metabolic signatures play a crucial role in the elevated risk of MASLD caused by air pollution. These results are relevant to patients and policymakers because they suggest that air pollution-related metabolic signatures are not only potentially associated with MASLD but also involved in mediating the process by which PM2.5, PM10, NO2, and NOx increase the risk of MASLD. Focusing on changes in air pollution-related metabolic signatures may offer a new perspective for preventing air pollution-induced MASLD and serve as protective measures to address this emerging public health challenge.

An Optimized Object Detection Algorithm for Marine Remote Sensing Images

In order to address the challenge of the small-scale, small-target, and complex scenes often encountered in offshore remote sensing image datasets, this paper employs an interpolation method to achieve super-resolution-assisted target detection. This approach aligns with the logic of popular GANs and generative diffusion networks in terms of super-resolution but is more lightweight. Additionally, the image count is expanded fivefold by supplementing the dataset with DOTA and data augmentation techniques. Framework-wise, based on the Faster R-CNN model, the combination of a residual backbone network and pyramid balancing structure enables our model to adapt to the characteristics of small-target scenarios. Moreover, the attention mechanism, random anchor re-selection strategy, and the strategy of replacing quantization operations with bilinear interpolation further enhance the model’s detection capability at a low cost. Ablation experiments and comparative experiments show that, with a simple backbone, the algorithm in this paper achieves a mAP of 71.2% on the dataset, an improvement in accuracy of about 10% compared to the Faster R-CNN algorithm.

PM2.5 chemical components are associated with in-hospital case fatality among acute myocardial infarction patients in China

Recent studies have linked the cardiovascular events with the exposure to ambient fine particulate matter (PM2.5); however, the impact of PM2.5 chemical components on acute myocardial infarction (AMI) case fatality remains poorly understood. To address this gap, we included 178,340 hospitalised patients with AMI utilising the inpatient discharge database from Sichuan, Shanxi, Guangxi, and Guangdong, China spanning 2014–2019. We evaluated exposure to PM2.5 and its components (black carbon (BC), organic matter (OM), sulphate (SO42−), nitrate (NO3−), and ammonium (NH4+)) using bilinear interpolation based on the patient’s residential address. We used mixed-effects logistic regression models to investigate the associations of PM2.5 and its five components with in-hospital AMI case fatality. Per interquartile range (IQR) increment in short-term exposure (7-day average) to overall PM2.5 (odds ratio (OR): 1.086, 95 % confidence interval (CI): 1.045–1.128), SO42−(1.063, 1.024–1.104), BC (1.055, 1.023–1.089), OM (1.052, 1.019–1.086, and NO3− (1.045, 1.003–1.089) were significantly associated with high risk of in-hospital AMI case fatality. The ORs per IQR increment in long-term exposure (annual average) were 1.323 (95 % CI: 1.255–1.394) for PM2.5, followed by BC (1.271, 1.210–1.335), OM (1.243, 1.188–1.300), SO42− (1.212, 1.157–1.270), NO3− (1.116, 1.075–1.159), and NH4+ (1.068, 1.031–1.106). Our study suggests that PM2.5 chemical components might be important risk factors for in-hospital AMI case fatality, highlighting the importance of targeted reduction of PM2.5 emissions, particularly BC, OM, and SO42−.

Ray-tracing versus Born approximation in full-sky weak lensing simulations of the MillenniumTNG project

ABSTRACT Weak gravitational lensing is a powerful tool for precision tests of cosmology. As the expected deflection angles are small, predictions based on non-linear N-body simulations are commonly computed with the Born approximation. Here, we examine this assumption using DORIAN, a newly developed full-sky ray-tracing scheme applied to high-resolution mass-shell outputs of the two largest simulations in the MillenniumTNG suite, each with a 3000 Mpc box containing almost 1.1 trillion cold dark matter particles in addition to 16.7 billion particles representing massive neutrinos. We examine simple two-point statistics like the angular power spectrum of the convergence field, as well as statistics sensitive to higher order correlations such as peak and minimum statistics, void statistics, and Minkowski functionals of the convergence maps. Overall, we find only small differences between the Born approximation and a full ray-tracing treatment. While these are negligibly small at power-spectrum level, some higher order statistics show more sizeable effects; ray-tracing is necessary to achieve per cent level precision. At the resolution reached here, full-sky maps with 0.8 billion pixels and an angular resolution of 0.43 arcmin, we find that interpolation accuracy can introduce appreciable errors in ray-tracing results. We therefore implemented an interpolation method based on non-uniform fast Fourier transforms (NUFFT) along with more traditional methods. Bilinear interpolation introduces significant smoothing, while nearest grid point sampling agrees well with NUFFT, at least for our fiducial source redshift, $z_s=1.0$, and for the 1 arcmin smoothing we use for higher order statistics.

Associations of ambient air pollution with incidence and dynamic progression of atrial fibrillation

The influence of air pollution on dynamic changes in clinical state from healthy to atrial fibrillation (AF), further AF-related complications and ultimately, death are unclear. We aimed to investigate the relationships between air pollution and the occurrence and progression trajectories of AF. We retrieved 442,150 participants free of heart failure (HF), myocardial infarction (MI), stroke and dementia at baseline from UK Biobank. Exposures to air pollution for each transition stage were estimated at the geocoded residential address of each participant using the bilinear interpolation approach. The outcomes were incident AF, complications, and death. Multi-stage models were used to evaluate the associations between air pollution and dynamic progression of AF. Over a 12.6-year median follow-up, a total of 21,670 incident AF patients were identified, of whom, 4103 developed complications and 1331 died. PM2.5, PM10, NOx and NO2 were differentially positively associated, while O3 was negatively associated with risks of progression trajectories of AF. PM2.5 exposure was significantly associated with an increased risk of progression. The associations of PM2.5, PM10, NOx, and NO2 on incident AF were generally more pronounced compared to other transitions. The cumulative transition probabilities were generally higher in individuals with higher exposure levels of PM2.5, PM10, NOx, and NO2 and lower exposure to O3. Air pollution could potentially have a role in increasing the risk of both the occurrence and progression of AF, emphasizing the significance of air pollution interventions in both the primary prevention of AF and the management of AF-related outcomes.

Generative adversarial networks with deep blind degradation powered terahertz ptychography

Ptychography is an imaging technique that uses the redundancy of information generated by the overlapping of adjacent light regions to calculate the relative phase of adjacent regions and reconstruct the image. In the terahertz domain, in order to make the ptychography technology better serve engineering applications, we propose a set of deep learning terahertz ptychography system that is easier to realize in engineering and plays an outstanding role. To address this issue, we propose to use a powerful deep blind degradation model which uses isotropic and anisotropic Gaussian kernels for random blurring, chooses the downsampling modes from nearest interpolation, bilinear interpolation, bicubic interpolation and down-up-sampling method, and introduces Gaussian noise, JPEG compression noise, and processed detector noise. Additionally, a random shuffle strategy is used to further expand the degradation space of the image. Using paired low/high resolution images generated by the deep blind degradation model, we trained a multi-layer residual network with residual scaling parameters and dense connection structure to achieve the neural network super-resolution of terahertz ptychography for the first time. We use two representative neural networks, SwinIR and RealESRGAN, to compare with our model. Experimental result shows that the proposed method achieved better accuracy and visual improvement than other terahertz ptychographic image super-resolution algorithms. Further quantitative calculation proved that our method has significant advantages in terahertz ptychographic image super-resolution, achieving a resolution of 33.09 dB on the peak signal-to-noise ratio (PSNR) index and 3.05 on the naturalness image quality estimator (NIQE) index. This efficient and engineered approach fills the gap in the improvement of terahertz ptychography by using neural networks.

Deep Convolutional Neural Network for Automated Staging of Periodontal Bone Loss Severity on Bite-wing Radiographs: An Eigen-CAM Explainability Mapping Approach.

Periodontal disease is a significant global oral health problem. Radiographic staging is critical in determining periodontitis severity and treatment requirements. This study aims to automatically stage periodontal bone loss using a deep learning approach using bite-wing images. A total of 1752 bite-wing images were used for the study. Radiological examinations were classified into 4 groups. Healthy (normal), no bone loss; stage I (mild destruction), bone loss in the coronal third (< 15%); stage II (moderate destruction), bone loss is in the coronal third and from 15 to 33% (15-33%); stage III-IV (severe destruction), bone loss extending from the middle third to the apical third with furcation destruction (> 33%). All images were converted to 512 × 400 dimensions using bilinear interpolation. The data was divided into 80% training validation and 20% testing. The classification module of the YOLOv8 deep learning model was used for the artificial intelligence-based classification of the images. Based on four class results, it was trained using fivefold cross-validation after transfer learning and fine tuning. After the training, 20% of test data, which the system had never seen, were analyzed using the artificial intelligence weights obtained in each cross-validation. Training and test results were calculated with average accuracy, precision, recall, and F1-score performance metrics. Test images were analyzed with Eigen-CAM explainability heat maps. In the classification of bite-wing images as healthy, mild destruction, moderate destruction, and severe destruction, training performance results were 86.100% accuracy, 84.790% precision, 82.350% recall, and 84.411% F1-score, and test performance results were 83.446% accuracy, 81.742% precision, 80.883% recall, and 81.090% F1-score. The deep learning model gave successful results in staging periodontal bone loss in bite-wing images. Classification scores were relatively high for normal (no bone loss) and severe bone loss in bite-wing images, as they are more clearly visible than mild and moderate damage.

Temperature Compensation Method Based on Bilinear Interpolation for Downhole High-Temperature Pressure Sensors.

Due to their high accuracy, excellent stability, minor size, and low cost, silicon piezoresistive pressure sensors are used to monitor downhole pressure under high-temperature, high-pressure conditions. However, due to silicon's temperature sensitivity, high and very varied downhole temperatures cause a significant bias in pressure measurement by the pressure sensor. The temperature coefficients differ from manufacturer to manufacturer and even vary from batch to batch within the same manufacturer. To ensure high accuracy and long-term stability for downhole pressure monitoring at high temperatures, this study proposes a temperature compensation method based on bilinear interpolation for piezoresistive pressure sensors under downhole high-temperature and high-pressure environments. A number of calibrations were performed with high-temperature co-calibration equipment to obtain the individual temperature characteristics of each sensor. Through the calibration, it was found that the output of the tested pressure measurement system is positively linear with pressure at the same temperatures and nearly negatively linear with temperature at the same pressures, which serves as the bias correction for the subsequent bilinear interpolation temperature compensation method. Based on this result, after least squares fitting and interpolating, a bilinear interpolation approach was introduced to compensate for temperature-induced pressure bias, which is easier to implement in a microcontroller (MCU). The test results show that the proposed method significantly improves the overall measurement accuracy of the tested sensor from 21.2% F.S. to 0.1% F.S. In addition, it reduces the MCU computational complexity of the compensation model, meeting the high accuracy demand for downhole pressure monitoring at high temperatures and pressures.

Enhanced recognition of insulator defects on power transmission lines via proposal-based detection model with integrated improvement methods

Deep learning-driven transmission line inspection is a critical area for smart power grid development. Despite advances in deep learning for insulator defect detection, challenges remain in model robustness and adaptability for the varying real-world adaptability, especially for insignificant defects in complex backgrounds. This study presents a comprehensive improvement strategy for detecting insulators and cross-scale broken defects on transmission lines, employing a proposal-based detection model. The model introduces a holistic pipeline of improved methods, including backbone modification, anchor box scale recalibration, and improvements in Region of Interest (RoI) downsampling alignment and Intersection over Union (IoU) loss function. Various backbone networks, including convolutional network (ConvNet) and Vision Transformer (ViT) structures, are constructed and integrated with attention modules, specifically designed to amplify the perception of insulators and defective regions. The geometric scale of anchor boxes is reconstructed using a developed clustering algorithm, considering the elongated characteristics of insulator strings to improve the adaptability of anchor boxes. Bilinear interpolation is utilized to mitigate spatial misalignment issues during the downsampling process of Region Proposal Network (RPN)-based proposals. The experimental results indicate that the improved models with the Swin Transformer (Swin-T) backbone framework achieve the mean Average Precision (mAP)@0.5 of 88.42% and mAP@0.7 of 60.52%, with a defect recall rate of 81.94%. Additionally, the improved IoU loss function contributes to the performance of the model at higher IoU thresholds. The results of this study contribute to the further development of defect detection frameworks for power vision applications.

Convolutional neural network advances in demosaicing for fluorescent cancer imaging with color-near-infrared sensors.

Single-chip imaging devices featuring vertically stacked photodiodes and pixelated spectral filters are advancing multi-dye imaging methods for cancer surgeries, though this innovation comes with a compromise in spatial resolution. To mitigate this drawback, we developed a deep convolutional neural network (CNN) aimed at demosaicing the color and near-infrared (NIR) channels, with its performance validated on both pre-clinical and clinical datasets. We introduce an optimized deep CNN designed for demosaicing both color and NIR images obtained using a hexachromatic imaging sensor. A residual CNN was fine-tuned and trained on a dataset of color images and subsequently assessed on a series of dual-channel, color, and NIR images to demonstrate its enhanced performance compared with traditional bilinear interpolation. Our optimized CNN for demosaicing color and NIR images achieves a reduction in the mean square error by 37% for color and 40% for NIR, respectively, and enhances the structural dissimilarity index by 37% across both imaging modalities in pre-clinical data. In clinical datasets, the network improves the mean square error by 35% in color images and 42% in NIR images while enhancing the structural dissimilarity index by 39% in both imaging modalities. We showcase enhancements in image resolution for both color and NIR modalities through the use of an optimized CNN tailored for a hexachromatic image sensor. With the ongoing advancements in graphics card computational power, our approach delivers significant improvements in resolution that are feasible for real-time execution in surgical environments.

Improving visual grounding with multi-modal interaction and auto-regressive vertex generation

We propose a concise and consistent network focusing on multi-task learning of Referring Expression Comprehension (REC) and Segmentation (RES) within Visual grounding (VG). To simplify the model architecture and achieve parameter sharing, we reconstruct the Visual grounding task as a floating-point coordinate generation problem based on both image and text inputs. Consequently, rather than separately predicting bounding boxes and pixel-level segmentation masks, we represent them uniformly as a sequence of coordinate tokens and output two corner points of bounding boxes and polygon vertices autoregressively. To improve the accuracy of point generation, we introduce a regression-based decoder. Inspired by bilinear interpolation, this decoder can directly predict precise floating-point coordinates, thus avoiding quantization errors. Additionally, we devise a Multi-Modal Interaction Fusion (M2IF) to address the imbalance between visual and language features in the model. This module focuses visual information on regions relevant to textual descriptions while suppressing the influence of irrelevant areas. Based on our model, Visual grounding is realized through a unified network structure. Experiments conducted on five benchmark datasets (RefCOCO, RefCOCO+, RefCOCOg, ReferItGame and Flickr30K Entities) demonstrate that the proposed unified network outperforms or is on par with many existing task-customized models. Codes are available at https://github.com/LFUSST/MMI-VG.

The robust SmartEnsembleNet: A game changer in finger knuckle biometrics

SmartEnsembleNet represents a major leap forward in the domain of biometric identification, focusing on the relatively underexplored area of finger knuckle biometrics. This research enhances the effectiveness and uniqueness of biometric systems by leveraging the distinct and secure features of finger knuckle patterns. As a complement to existing biometric techniques like fingerprints and facial recognition, it adds an additional layer of security and reliability. Addressing the challenges faced by finger knuckle identification, such as image acquisition, environmental effects, and aging, SmartEnsembleNet introduces a pioneering multi-layered ensemble learning strategy that combines eight diverse machine learning models. These models, trained via a 5-fold cross-validation on datasets from IIT Delhi and PolyU, demonstrate significant improvements in accuracy, precision, recall, and F1-score. A distinctive feature of SmartEnsembleNet is its innovative meta-model, streamlining the creation of a meta-dataset. This dataset is created by applying bilinear interpolation to the predicted probabilities from the top-performing model, and it is further refined by subtracting these interpolated probabilities from the validation data. This iterative process generates ‘smart data’, which serves as the basis for training the ultimate ‘smart model’. Logistic regression is chosen for its outstanding performance. This method, applied uniformly to both training and testing phases, considerably enhances classification accuracy in finger knuckle biometrics, establishing new benchmarks. SmartEnsembleNet’s performance undergoes a detailed evaluation using the Friedman test and Nemenyi post-hoc analysis, guaranteeing a selection process grounded in statistical rigor. With its resilience to noise and adaptability under various conditions, SmartEnsembleNet not only addresses existing gaps in finger knuckle biometric research but also pioneers future advancements, providing a versatile and secure solution impactful for both academia and high-security scenarios.

Evaluation of extra pixel interpolation with mask processing for medical image segmentation with deep learning

Current mask processing operations rely on interpolation algorithms that do not produce extra pixels, such as nearest neighbor (NN) interpolation, as opposed to algorithms that do produce extra pixels, like bicubic (BIC) or bilinear (BIL) interpolation. In our previous study, the author proposed an alternative approach to NN-based mask processing and evaluated its effects on deep learning training outcomes. In this study, the author evaluated the effects of both BIC-based image and mask processing and BIC-and-NN-based image and mask processing versus NN-based image and mask processing. The evaluation revealed that the BIC-BIC model/network was an 8.9578% (with image size 256 × 256) and a 1.0496% (with image size 384 × 384) increase of the NN-NN network compared to the NN-BIC network which was an 8.3127% (with image size 256 × 256) and a 0.2887% (with image size 384 × 384) increase of the NN-NN network.

Path planning of PRM based on artificial potential field in radiation environments

The radiation environment in nuclear facility work sites is complex. To ensure the safety of personnel working in these environments, it is essential to develop an optimal path planning method based on the criterion of personnel radiation dose. In this paper an improved probabilistic roadmap algorithm (APF-PRM) is proposed based on the combination of artificial potential fields (APF) and probabilistic roadmap algorithm (PRM) for path planning in complex static radiation environments. The entire two-dimensional dose field model of the radiation environment is constructed using bilinear interpolation based on simulation data from Geant4. Using this dose field model and the personnel radiation dose calculation model, the new algorithm is integrated with the A* algorithm to perform path planning in the radiation environment. To assess the performance of the APF-PRM algorithm, a comparison with other algorithms is made in four cases. In the first case, the computation speed of APF-PRM algorithm is significantly faster than that of the conventional A* algorithm in the same environment. In case 2, a reduction of 16.25% in cumulative personnel radiation dose is observed when compared with the traditional PRM algorithm, and the results from multiple repetitions of the experiment display higher stability. In case 3 a comparison between the APF-PRM algorithm and the Voronoi diagram is made, revealing a 36.92% reduction in cumulative dose, along with a 16.49% decrease in path length. In case 4, the four algorithms, A*, APF-PRM, PRM, and Voronoi, are comprehensively compared under the same circumstances, and the parameters of the APF-PRM algorithm that meets the requirements of minimum dose and optimal path are found.

An improved dynamic bidirectional coupled hydrologic–hydrodynamic model for efficient flood inundation prediction

Abstract. To improve computational efficiency while maintaining numerical accuracy, coupled hydrologic–hydrodynamic models based on non-uniform grids are used for flood inundation prediction. In these models, a hydrodynamic model using a fine grid can be applied to flood-prone areas, and a hydrologic model using a coarse grid can be used for the remaining areas. However, it is challenging to deal with the separation and interface between the two types of areas because the boundaries of the flood-prone areas are time dependent. We present an improved Multigrid Dynamical Bidirectional Coupled hydrologic–hydrodynamic Model (IM-DBCM) with two major improvements: (1) automated non-uniform mesh generation based on the D-infinity algorithm was implemented to identify the flood-prone areas where high-resolution inundation conditions are needed and (2) ghost cells and bilinear interpolation were implemented to improve numerical accuracy in interpolating variables between the coarse and fine grids. A hydrologic model, the 2D nonlinear reservoir model, was bidirectionally coupled with a 2D hydrodynamic model that solves the shallow-water equations. Three cases were considered to demonstrate the effectiveness of the improvements. In all cases, the mesh generation algorithm was shown to efficiently and successfully generate high-resolution grids in those flood-prone areas. Compared to the original M-DBCM (OM-DBCM), the new model had lower root-mean square errors (RMSEs) and higher Nash–Sutcliffe efficiencies (NSEs), indicating that the proposed mesh generation and interpolation were reliable and stable. It can be adequately adapted to the real-life flood evolution process in watersheds and provide practical and reliable solutions for rapid flood prediction.

Reconstructing Cancellous Bone From Down-Sampled Optical-Resolution Photoacoustic Microscopy Images With Deep Learning

ObjectiveBone diseases deteriorate the microstructure of bone tissue. Optical-resolution photoacoustic microscopy (OR-PAM) enables high spatial resolution of imaging bone tissues. However, the spatiotemporal trade-off limits the application of OR-PAM. The purpose of this study was to improve the quality of OR-PAM images without sacrificing temporal resolution. MethodsIn this study, we proposed the Photoacoustic Dense Attention U-Net (PADA U-Net) model, which was used for reconstructing full-scanning images from under-sampled images. Thereby, this approach breaks the trade-off between imaging speed and spatial resolution. ResultsThe proposed method was validated on resolution test targets and bovine cancellous bone samples to demonstrate the capability of PADA U-Net in recovering full-scanning images from under-sampled OR-PAM images. With a down-sampling ratio of [4, 1], compared to bilinear interpolation, the Peak Signal-to-Noise Ratio and Structural Similarity Index Measure values (averaged over the test set of bovine cancellous bone) of the PADA U-Net were improved by 2.325 dB and 0.117, respectively. ConclusionThe results demonstrate that the PADA U-Net model reconstructed the OR-PAM images well with different levels of sparsity. Our proposed method can further facilitate early diagnosis and treatment of bone diseases using OR-PAM.

Robot Grasp Detection with Loss-Guided Collaborative Attention Mechanism and Multi-Scale Feature Fusion

Grasp detection serves as the fundamental element for achieving successful grasping in robotic systems. The encoder–decoder structure has become widely adopted as the foundational architecture for grasp detection networks due to its inherent advantages of speed and accuracy. However, traditional network structures fail to effectively extract the essential features required for accurate grasping poses and neglect to eliminate the checkerboard artifacts caused by inversion convolution during decoding. Aiming at overcoming these challenges, we propose a novel generative grasp detection network (LGAR-Net2). A transposed convolution layer is employed to replace the bilinear interpolation layer in the decoder to remove the issue of uneven overlapping and consequently eliminate checkerboard artifacts. In addition, a loss-guided collaborative attention block (LGCA), which combines attention blocks with spatial pyramid blocks to enhance the attention to important regions of the image, is constructed to enhance the accuracy of information extraction. Validated on the Cornell public dataset using RGB images as the input, LGAR-Net2 achieves an accuracy of 97.7%, an improvement of 1.1% over the baseline network, and processes a single RGB image in just 15 ms.

An enhanced approach for few-shot segmentation via smooth downsampling mask and label smoothing loss

Few-shot semantic segmentation aims to segment new categories with only a small number of annotated images. Previous methods mainly focused on exploiting the pixel-level correlation between the support image and the query image, combined with attention-based methods, resulting in significant advancements. In this paper, we introduce a new perspective to enhance few-shot segmentation. We identify that utilizing the bilinear interpolation method to downsample the mask leads to the loss of fine-grained information from the target features. To address this issue, we propose a Smooth Downsampling Mask (SDM) method. The SDM method is designed to retain more effective target semantic features by employing a cascaded downsampling approach with a smooth kernel for mask processing. Additionally, we propose a label smoothing loss to further enhance the performance, which provides direct guidance for low-resolution feature map optimization. Both methods can be used as plug-and-play modules for existing methods. Notably, our proposed method does not involve additional learnable parameters and is computationally efficient, thus achieving painless gains. To validate the effectiveness of our method, we take three publicly available models as baselines and conduct extensive experiments on three public benchmarks PASCAL-5i, COCO-20i and FSS-1000, and achieve considerable improvement.