Retrieval Method Research Articles

Enhanced sperm isolation via bulk acoustic waves for high-throughput motility screening

The increasing infertility rate has become a worrying global challenge in recent years. According to the report of the World Health Organization, the male factor is responsible for over half of infertility cases, which includes the lack of desirable characteristics in sperm motility, morphology, and DNA integrity. In recent years, it has been shown that clinical methods including density gradient centrifugation cause damage to sperm DNA and besides being invasive, they are costly and time-consuming. In contrast, microfluidics has been used as a promising and non-invasive approach to manipulate biological cells. Here, by using the microvortices created by the oscillation of the bubbles caused by the bulk acoustic waves, we were able to trap sperms with less motility. In contrast, the highly motile sperms overcame the force of the microvortices and were guided to the outlet pool by following the channel boundaries. As a result, over 50% and 44% improvement in sperm progressive motility and viability, respectively, as well as 40% improvement in DNA integrity, were observed in the analysis of sperms retrieved from the output pool. In addition to being fast and non-invasive, the proposed device benefits from an easy method for sperms retrieval and does not require any preprocessing of the raw sperm sample.

Linking Vegetation Phenology to Net Ecosystem Productivity: Climate Change Impacts in the Northern Hemisphere Using Satellite Data

With global climate change, linking vegetation phenology with net ecosystem productivity (NEP) is crucial for assessing vegetation carbon storage capacity and predicting terrestrial ecosystem changes. However, there have been few studies investigating the relationship between vegetation phenology and NEP in the middle and high latitudes of the Northern Hemisphere. This study comprehensively analyzed vegetation phenological changes and their climate drivers using satellite data. It also investigated the spatial distribution and climate drivers of NEP and further analyzed the sensitivity of NEP to vegetation phenology. The results indicated that the average land surface phenology (LSP) was dominated by a monotonic trend in the study area. LSP derived from different satellite products and retrieval methods exhibited relatively consistent responses to climate. The average SOS and POS for different retrieval methods showed a higher negative correlation with nighttime temperatures compared to daytime temperatures. The average EOS exhibited a higher negative correlation with daytime temperatures than a positive correlation. The correlations between VPD and the average SOS, POS, and EOS showed that the proportion of negative correlations was higher than that of positive correlations. The average annual NEP ranged from 0 to 1000 gC·m−2. The cumulative trends of NEP were mainly monotonically increasing, accounting for 61.04%, followed by monotonically decreasing trends, which accounted for 17.95%. In high-latitude regions, the proportion of positive correlation between VPD and NEP was predominant, while the proportion of negative correlation was predominant in middle-latitude regions. The positive and negative correlations between soil moisture and NEP (48.08% vs. 51.92%) were basically consistent in the study area. The correlation between SOS and POS with NEP was predominantly negative. The correlation between EOS and NEP was overall characterized by a greater proportion of negative correlations than positive correlations. The correlation between LOS and NEP exhibited a positive relationship in most areas. The sensitivity of NEP to vegetation phenological parameters (SOS, POS, and EOS) was negative, while the sensitivity of NEP to LOS was positive (0.75 gC·m−2/d for EVI vs. 0.63 gC·m−2/d for LAI vs. 0.30 gC·m−2/d for SIF). This study provides new insights and a theoretical basis for exploring the relationship between vegetation phenology and NEP under global climate change.

Text-video retrieval re-ranking via multi-grained cross attention and frozen image encoders

State-of-the-art methods for text-video retrieval generally leverage CLIP embeddings and cosine similarity for efficient retrieval. Meanwhile, recent advancements in cross-attention techniques introduce transformer decoders to facilitate attention computation between text queries and visual tokens extracted from video frames, enabling a more comprehensive interaction between textual and visual information. In this study, we combine the advantages of both approaches and propose a fine-grained re-ranking approach incorporating a multi-grained text-video cross attention module. Specifically, the re-ranker enhances the top K similar candidates identified by the cosine similarity network. To explore video and text interactions efficiently, we introduce frame and video token selectors to obtain salient visual tokens at both frame and video levels. Then, a multi-grained cross-attention mechanism is applied between text and visual tokens at these levels to capture multimodal information. To reduce the training overhead associated with the multi-grained cross-attention module, we freeze the vision backbone and only train the multi-grained cross attention module. This frozen strategy allows for scalability to larger pre-trained vision models such as ViT-G, leading to enhanced retrieval performance. Experimental evaluations on text-video retrieval datasets showcase the effectiveness and scalability of our proposed re-ranker combined with existing state-of-the-art methodologies.

Measurements of microwave transmission mode in HL-3 tokamak relevant ECH corrugated waveguides transmission line

The transmission modes of high-power millimeter wave in corrugated waveguides transmission line (TL) are crucial, especially for the long-pulse operation for long-distance TL. The microwave transmission mode in the TL on HL-3 tokamak relevant the electron cyclotron heating system are studied for the purpose according to the phase retrieval method based on intensity distribution measured by an infrared camera at several different locations away from the outlet of TL. A specific high power test platform with about 40-meter TL is established and the mode contents in the middle and the end of the TL are analyzed. In the middle of the TL, the proportion of main transmission mode LP01 mode is about 97% and other high-order modes contents are less than 3%. In the end of the TL, the main mode purity is about 94%, and the high-order mode, LP11 even mode is excited about 4.6%. These mode analysis results indicate that the high-power and long-pulse millimeter wave corrugated waveguides TL are designed and machined well, and the alignment method for long distance transmission line installation is highly useful for high-purity wave mode transmission.

Philosophical analysis of Mortimer J. Adler's Christian education and global education management

The present research shows the absurdity of education, especially education management, in the posthumanist schema. Posthumanists seem to want to form their own group, but researchers see that posthumanism is not new, especially in the world of education. Philosopher Mortimer J. Adler, in the framework of perennialism, shows the structure of education, especially in America, which, according to researchers, is relevant in Indonesia. Research findings are based on qualitative research with data retrieval methods from literature studies. Results showed that Mortimer J. Adler's philosophical thinking emphasizes the expansion of education management in both depth and fragility by and through learners, parents, teachers and related parties in the mecca of an independent curriculum.

Foreign bodies in the ear and nose of the Australian paediatric population: A standardised approach.

Foreign bodies (FBs) in the ears and noseare common presentations among paediatric patients, necessitating prompt and appropriate management to avoid potential complications. This article presents a standardised approach for FB retrieval based on local clinical data from a tertiary paediatric centre, addressing the challenges of identifying patients requiring specialist referral. The approach involves a detailed initial consultation and succinct examination todetermine the most suitable retrieval method. Patient preparation and equipment availability are key to successful outcomes. A combination of microsuction and alligator forceps with adequate illumination is strongly recommended, along with various adjunct tools commonly available. Urgent referral to otorhinolaryngology is indicated for button batteries and previous failed attempts. This comprehensive approach seeks to achieve successful FB removal and optimise patient outcomes in both community and hospital settings.

Comparisons for Different State-of-art Planet Searching Scenarios: Transit, Radial Velocity and Direct-imaging

The vastness of the universe has inspired mankind’s infinite imagination of the exoplanets with living. The exploration of them reflects human beings’ yearning for possible life and the pursuit of a possible future home. From the first detection of an exoplanet in 1992 to the detection of 5,747 exoplanets on August 16, 2024, scholars have worked hard to explore. This study presents typical results of exoplanet detection in recent years, describing in general terms which planet retrieval methods have been used, what results are obtained by each method and which parameters of the planet can be detected. To be specific, three main methods for detecting exoplanets are detailly evaluated, namely the transit, the radial velocity and the direct imaging. By comparing these three methods, detection methods suitable for different types of exoplanets are summarized. According to the analysis, the limitations of current exoplanet detection are summarized based on the results, and future countermeasures are outlined. These results offer a guideline for exoplanet exploration and shed light on understanding the three main methods of detecting exoplanets.

An enhanced method for reconstruction of full SIF spectrum for near-ground measurements

Recently the applications of remotely sensed Solar-Induced chlorophyll Fluorescence (SIF) in the study of photosynthesis, stress conditions, and gross primary production have increased significantly. The full SIF spectrum spans over a spectral region of 650 ∼ 850 nm with two characteristic peaks around 685 nm and 740 nm. Over recent decades, many retrieval algorithms have been developed to estimate SIF at Top-Of-Canopy (TOC) using in-situ measurements of solar irradiance and canopy radiance spectra. Although the majority of retrieval methods retrieve SIF at a narrow spectral window, there exists a potential for retrieval of SIF in the full emission spectrum. Moreover, solar irradiance and canopy radiance spectra should ideally be measured at the same time but are usually measured sequentially with a certain time lag, raising potential errors in SIF retrieval. In this study, an enhanced retrieval algorithm of the full SIF spectrum at TOC is proposed. The proposed algorithm attempts to minimize the errors owing to time mismatch in measurements of solar irradiance and canopy radiance spectra. As an improvement to the previous algorithm (advanced Fluorescence Spectrum Reconstruction, aFSR), this proposed algorithm (aFSR-SVE) models the SIF-free contribution with principal components using the singular value decomposition technique. The optimal parameter settings in the forward model were determined for the experimental data collected by spectrometers used in the study. Firstly, the proposed algorithm was used to reconstruct full SIF spectrum for simulated data. The results were compared with known reference SIF values. After achieving satisfying results from simulated data, the proposed algorithm was compared with retrievals from established algorithms using experimental data. The results show improved SIF retrieval accuracy, without the need to simultaneously measure solar irradiance and canopy radiance spectra. The retrieval values comply with the results of previous algorithms in terms of spectral shape, diurnal trend, and temporal variations. The induced errors in SIF retrievals due to non-simultaneous measurements of solar irradiance and canopy radiance spectra were also investigated and the proposed algorithm was found to be less prone to such errors. Hence, the proposed algorithm is an improvement in reconstructing the full SIF spectrum with near-ground measurements. With the help of the proposed algorithm, field measurements using sequential systems and automated measurements of multiple targets can be performed effectively as it relaxes the requirement of concurrent measurement of solar irradiance and canopy radiance spectra. For future work, the applicability of this method can be investigated under more variable illumination conditions, like high cirrus clouds, passing clouds or persistent thin clouds.

Precipitation Retrieval from FY-3G/MWRI-RM Based on SMOTE-LGBM

Using the FY-3G/MWRI-RM observations, this paper proposes a precipitation retrieval method that combines the Synthetic Minority Over-sampling Technique with Light Gradient Boosting Machine (SMOTE-LGBM) and analyzes the impact of MWRI-RM channel settings on precipitation retrieval. The SMOTE-LGBM-based model consists of two LGBM models for precipitation identification and estimation, respectively. The SMOTE method is used to address the imbalance between precipitation and non-precipitation samples. Using the Integrated Multi-Satellite Retrievals for the Global Precipitation Measurement (IMERG) product as a reference, we validate the retrieved precipitation by the SMOTE-LGBM-based model with an independent testing dataset. The critical success indexes are 0.483 and 0.526, and the Pearson correlation coefficients are 0.611 and 0.645 for the ocean and land regions, respectively. The spatial distributions of the retrieved and IMERG accumulated precipitation in the testing dataset are similar. In addition, we visualize and analyze the cases of Meiyu and two typhoons. The results indicate that the SMOTE-LGBM-based model effectively represents the spatial distribution characteristics of precipitation and achieves high agreement with IMERG precipitation products. Overall, the SMOTE-LGBM-based model successfully retrieves precipitation from MWRI-RM and provides accurate precipitation products for FY-3G/MWRI-RM for the first time.

Development of an Evaluation Instrument on Artificial Intelligence Search Tools for Evidence Synthesis

What Was the Question? There are conflicting calls for evidence synthesis producers to adopt or avoid recent Artificial Intelligence (AI) technologies. Our question was: How do we evaluate rapidly evolving AI tools to enhance the production of evidence syntheses and maintain quality standards? What Did We Do? To advance information retrieval science for producing evidence syntheses at Canada’s Drug Agency (CDA-AMC), the Research Information Services team developed a process to evaluate promising AI search tools. We inventoried 51 tools in the fall of 2023, established selection criteria, assessed specific attributes, and built a standalone instrument to support continuous monitoring and evaluation of new tools. What Did We Find? Rapid development of AI search tools requires a flexible evaluation instrument to inform adoption decisions and enable comparison between tools. We identified mandatory and desirable characteristics for suitable AI tools to assist with information retrieval tasks conducted by CDA-AMC. This work enabled the development of a flexible instrument to evaluate novel AI search tools for evidence synthesis. What Does This Mean? CDA-AMC operationalized a replicable process to monitor and evaluate AI search tools. Our approach to evaluating AI technologies will advance information retrieval methods by CDA-AMC, and our evaluation instrument will assist any evidence synthesis producer interested in adopting AI search tools.

Intercomparison of aerosol optical depth retrievals from GAW-PFR and SKYNET sun photometer networks and the effect of calibration

Abstract. In this study, we assess the homogeneity of aerosol optical depth (AOD) between two sun photometer networks, the Global Atmosphere Watch-Precision Filter Radiometer (GAW-PFR) and the European SKYNET radiometers network (ESR), at the common wavelengths of their main instruments (500 and 870 nm). The main focus of this work is to evaluate the effect of the improved Langley plot calibration method (ILP) used by SKYNET and to investigate the factors affecting its performance. We used data from three intercomparison campaigns that took place during 2017–2021. Each campaign was organized at two locations (mountainous rural – Davos, Switzerland; urban – Rome, Italy). Our analysis shows that differences in AOD due to post-processing and instrument differences are minor. The main factor leading to AOD differences is the calibration method. We found a systematic underestimation of AOD in ESR compared to in GAW-PFR due to underestimation of the calibration constant calculated with the ILP method compared to the calibration transfers using the PFR as a reference. The calibration and AOD differences are smaller in Davos, where the traceability criteria are satisfied at 870 nm and where the median differences are below 0.01 at 500 nm. In Rome, the AOD median differences at 500 nm were in the 0.015–0.034 range. We conducted a sensitivity study, which shows that part of the difference can potentially be explained by errors in the assumed surface albedo and instrument solid-view angle provided as inputs to the ILP code (based on Skyrad pack 4.2). Our findings suggest that the ILP method is mainly sensitive to the measured sky radiance. The underestimation in calibration is probably caused by an error in the retrieved scattering AOD (sc-AOD) through the sky radiance inversion. Using an alternative retrieval method (Skyrad MRI pack version 2) to derive sc-AOD and to recalibrate the instruments with the ILP method, we found no significant differences between the retrieved sc-AOD and no systematic increase in the ILP-derived calibration constant when using the MRI pack for sc-AOD inversion instead of the Skyrad 4.2. The potential error may be a result of the model assumptions used for the sky radiance simulations. In conclusion, the on-site calibration of sun photometers has several advantages, including the fact that instrument shipments and data gaps can be avoided. However, it has also the disadvantages of a larger uncertainty and significant systematic differences compared to the traditional Langley calibration performed under low- and constant-AOD conditions at high-altitude sites. The larger uncertainty of the ILP method can be attributed to the required modelling and input parameters.

The comparison between single-point method and footprint-integrated validation method of the remote-sensing retrieval of evapotranspiration: a case study at Daman site

ABSTRACT In single-site validations, site-pixel, synthetic pixel and footprint-integrated methods are commonly used to validate evapotranspiration (ET) retrieval at the pixel scale. To reveal the reliability of these methods, this study analysed the performance of the three methods for ET retrieval validation based on eddy-covariance observations at the Daman site which is numerously used for validation in Heihe Watershed Allied Telemetry Experimental Research (HiWATER). The analysis at different spatial scales in different seasons showed that the validation results of the site-pixel method were more similar that of the footprint-integrated method than the synthetic pixel method. However, the validation results of the site-pixel method showed significant discrepancies from those of the footprint-integrated method, especially in summer at the 480 M scale which is near to the spatial resolution of MODIS with a large difference of more than 125.65 W M−2. This error is due to aggregation and high spatial heterogeneity. The use of synthetic pixel can avoid this error. Normally, the error probably appears when the wind direction is 270°–360°, usually in summer, autumn and winter. In addition, the different aggregation methods had a slight effect on the ET estimation results. This study is helpful in determining the reliability of different single-point validation methods and provides a reliable way to select a validation method for single-site validation of remote sensing retrieval of ET.

Remote sensing insights into water allocation and evaporation challenges in the Hirmand River Basin, after the operation of Kamal Khan Dam

Study RegionThe Hirmand River Basin is a vital transboundary river system, that originates in Afghanistan’s Hindu Kush Mountains and flows into the Sistan Depression, and encompassing the Chah Nimeh Reservoirs in Iran and the Godzareh Depression in Afghanistan. Study FocusThe Kamal Khan Dam, constructed on the Hirmand River in Afghanistan, has significantly altered the downstream water direction and distribution between the Chah Nimeh Reservoirs and Godzareh Depression. Utilizing remote sensing techniques, particularly Landsat 8 satellite imagery and the FAO 56 PM as a evaporation retrieval method, the research focuses on evaluating changes in water allocation and evaporation rates in these regions over the past decade. New Hydrological Insights for the RegionThe findings reveal that after operation of the Kamal Khan Dam, water allocation to the Chah Nimeh Reservoirs has drastically decreased, leading to a 54 % reduction in their average area from 2020–2023 compared to the previous years. Conversely, the Godzareh Depression, now receiving the redirected water, has experienced significantly higher evaporation rates, contributing to substantial water losses. These changes underscore the critical need for effective water management strategies to address the escalating water scarcity and hydrological imbalances in this arid region.

Future prospects for the advancement of treatment of men with NOA: focus on gene editing, artificial sperm, stem cells, and use of imaging.

Nonobstructive azoospermia (NOA) affects about 60% of men with azoospermia, representing a severe form of male infertility. The current approach to manage NOA primarily involves testicular sperm retrieval methods such as conventional testicular sperm extraction (c-TESE) and microdissection testicular sperm extraction (micro-TESE). While combining testicular sperm retrieval with intracytoplasmic sperm injection (ICSI) offers hope for patients, the overall sperm retrieval rate (SRR) stands at around 50%. In cases where micro-TESE fails to retrieve sperm, limited options, like donor sperm or adoption, can be problematic in certain cultural contexts. This paper delves into prospective treatments for NOA management. Gene editing technologies, particularly clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated (Cas) protein 9 (CRISPR/Cas9), hold potential for correcting genetic mutations underlying testicular dysfunction. However, these technologies face challenges due to their complexity, potential off-target effects, ethical concerns, and affordability. This calls for research to address key challenges associated with NOA management within the clinical settings. This also necessitate ongoing research essential for developing more sensitive diagnostic tests, validating novel treatments, and customizing current treatment strategies for individual patients. This review concluded that the future of NOA management may entail a combination of these treatment options, tailored to each patient's unique circumstances, providing a comprehensive approach to address NOA challenges.

High-resolution lensless holographic microscopy using a physics-aware deep network.

Lensless digital inline holographic microscopy (LDIHM) is an emerging quantitative phase imaging modality that uses advanced computational methods for phase retrieval from the interference pattern. The existing end-to-end deep networks require a large training dataset with sufficient diversity to achieve high-fidelity hologram reconstruction. To mitigate this data requirement problem, physics-aware deep networks integrate the physics of holography in the loss function to reconstruct complex objects without needing prior training. However, the data fidelity term measures the data consistency with a single low-resolution hologram without any external regularization, which results in a low performance on complex biological data. We aim to mitigate the challenges with trained and physics-aware untrained deep networks separately and combine the benefits of both methods for high-resolution phase recovery from a single low-resolution hologram in LDIHM. We propose a hybrid deep framework (HDPhysNet) using a plug-and-play method that blends the benefits of trained and untrained deep models for phase recovery in LDIHM. The high-resolution phase is generated by a pre-trained high-definition generative adversarial network (HDGAN) from a single low-resolution hologram. The generated phase is then plugged into the loss function of a physics-aware untrained deep network to regulate the complex object reconstruction process. Simulation results show that the SSIM of the proposed method is increased by 0.07 over the trained and 0.04 over the untrained deep networks. The average phase-SNR is elevated by 8.2dB over trained deep models and 9.8dB over untrained deep networks on the experimental biological cells (cervical cells and red blood cells). We showed improved performance of the HDPhysNet against the unknown perturbation in the imaging parameters such as the propagation distance, the wavelength of the illuminating source, and the imaging sample compared with the trained network (HDGAN). LDIHM, combined with HDPhysNet, is a portable and technology-driven microscopy best suited for point-of-care cytology applications.

Testing a spaceborne passive‐microwave severe hail retrieval over Argentina using ground‐based dual‐polarization radar data

AbstractPassive‐microwave hail detection methods rely on the scattering signal produced by large ice hydrometeors. Recently, passive‐microwave hail retrieval methods have been trained by pairing Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and Global Precipitation Measurement (GPM) Microwave Imager (GMI) data with United States surface hail reports to present a global hail climatology. In particular, Southeastern South America (SESA) stands out as a hotspot for hail occurrence, more specifically northeastern Argentina. However, the analysis performed with the available ground hail reports in Argentina shows some differences from the available ground hail reports in Argentina, as it locates the most frequent hail storms in the southwestern (Mendoza and Córdoba) and northeastern (Misiones) regions of SESA. To assess the retrieval from a regional perspective, we take a case study approach and focus on polarization‐corrected temperature features (PCTFs) with a detected . We mainly rely on polarimetric weather radars and a hydrometeor identification algorithm (HID) as a proxy for ground truth. Results indicate that the hail retrieval responds mainly to low 37‐GHz polarized corrected (brightness) temperatures (PCTs) and large PCTF area, while HID outputs show that high probabilities of hail also correspond to abundantly suspended graupel. The observed passive‐microwave hail detection hotspot could be influenced by mesoscale convective systems with deep columns of graupel‐sized frozen hydrometeors, common in northeastern Argentina.

Retrieval of total suspended matter concentration in the yellow river estuary offshore area based on QAA-RF model

ABSTRACT Total suspended matter is one of the crucial water quality parameters for both inland and marine environments, and a key role in evaluating the water quality of estuaries and offshore areas. Each year, the Yellow River carries a significant amount of sediment into the semi-enclosed Bohai Sea, results in a prolonged high concentration of total suspended matter in the offshore areas of the Yellow River Estuary. This study focuses on the offshore region of the Yellow River Estuary in China. Utilizing Sentinel-2 satellite imagery data from 2020 to 2023 and in-situ measured data from August 2020 to August 2022, to address the lack of physical mechanisms currently studied in machine learning retrieval methods, a model that integrates the physics-driven Quasi-Analytical Algorithm (QAA) and data-driven Random Forest (RF) is employed for the retrieval of total suspended matter concentration in the study area. The fused model (QAA-RF) is compared and analysed against regression models and standalone machine learning models. The results indicate that the accuracy of machine learning retrieval models is consistently higher than that regression models. The QAA-RF model demonstrates the highest accuracy (R 2 = 0.87, MAE = 5.01 mg L−1, RMSE = 6.39 mg L−1). Based on the QAA-RF model and the imagery data, monthly total suspended matter concentration is conducted in the study area. The retrieval results indicates that: (1) the total suspended matter concentrations is primarily concentrated in the near estuary region, with concentrations decreasing as distance from the estuary increases. (2) the total suspended matter concentration exhibits a distribution pattern with higher values in spring and winter, and lower values in summer and autumn. (3) the concentration of total suspended matter shows relatively small fluctuations at the annual scale from 2020 to 2023.

Analysis of the underwater acoustic environment in the Catalan Shore: Badalona and Port of Barcelona cases in the framework of the Deuteronoise JPI-Oceans Project

Maritime traffic noise pollution is a significant concern for ocean's health and living organisms' welfare, particularly in European sea basins. Research has predominantly focused on vertebrates, linking noise pollution to hearing impairment, affecting their predator and prey detection, communication, and navigation. However, recent findings suggest that tunicates, marine invertebrates related to vertebrates, may also be susceptible to noise due to their similar acoustic and particle motion sensory organs. The JPI Oceans project, Deuteronoise, addresses this research gap by characterizing noise from maritime traffic in selected European sites, investigating the effects of noise on marine deuterostomes. In this paper, we introduce the pipeline and retrieval methods, followed by an analysis of the data collected from two contrasting locations: Port of Barcelona, noise polluted area, and Badalona, non-polluted one, focusing on the noise impact on Oikopleura dioica, a representative appendicularian tunicate. Initial laboratory analyses of recorded sounds have identified different vessel and sound features, providing distinction between the two sites. With this data analysis, we aim to bring valuable insights that will enhance our understanding of maritime noise pollution. This research not only advances our knowledge of the ecological consequences of noise pollution but also informs conservation strategies for vulnerable marine ecosystems.

Global-aware Fragment Representation Aggregation Network for image-text retrieval

Image-text retrieval is an important kind of cross-modal retrieval method and has recently attracted much attention. Existing image-text retrieval methods often ignore the relative importance of each fragment (region in an image or word in a sentence) on the global semantic of image or text when aggregating features of image or text fragments, resulting in the ineffectiveness of the learned image and text representations. To address this problem, we propose an image-text retrieval method named Global-aware Fragment Representation Aggregation Network (GFRAN). Specifically, it first designs a fine-grained multimodal information interaction module based on the self-attention mechanism to model both the intra-modality and inter-modality relationships between image regions and words. Then, with the guidance of the global image or text feature, it aggregates image or text fragment features conditioned on their attention weights over global feature, to highlight fragments that contribute more to the overall semantics of images and texts. Extensive experiments on two benchmark datasets Flickr30K and MS-COCO demonstrate the superiority of the proposed GFRAN model over several state-of-the-art baselines.

Method of generating potential evapotranspiration with high precision and resolution

Potential evapotranspiration (PET) is a key factor in hydrological cycle and energy balance and plays an important role in drought and global climate change response. Existing observational and modeling methods for PET retrieval have their limitations, such as low precision and poor spatial resolution, which becomes the focus of this study. A hybrid PET fusion (HPF) method is proposed by fusing station- and grid-based PET, in which the PET expression is determined by considering the factors of location, temperature, and zenith total delay (ZTD). In addition, an improved Helmert variance component estimation method is introduced to determine the optimal weights of the HPF model. Corresponding data, which include monthly Thornthwaite (TH)-derived PET data with a spatial resolution of 0.25° × 0.25° and Penman–Monteith (PM)-derived PET data at 704 meteorological stations, over the past 60 years from 1959 to 2018 in China are selected. The 10-fold cross-validation method is introduced to evaluate the internal and external accuracies of the proposed HPF method. Statistical result shows that the average root mean square (RMS) of the proposed HPF method is 13.98 mm, with an average RMS improvement rate (IR) of 46.71 % compared with TH-derived PET, when PM-derived PET is regarded as a reference. Moreover, the performance of the HPF-derived standardized precipitation evapotranspiration index (SPEI) is evaluated at different time scales, and the average RMS is 0.3, with an average RMS IR of 26.33 % compared with TH-derived SPEI. Such results verify the good performance of the proposed HPF model and enrich the methods for obtaining PET with high precision and resolution.