Multidimensional Measure Matching for Crowd Counting.

This article addresses the challenge of scale variations in crowd-counting problems from a multidimensional measure-theoretic perspective. We start by formulating crowd counting as a measure-matching problem, based on the assumption that discrete measures can express the scattered ground truth and the predicted density map. In this context, we introduce the Sinkhorn counting loss and extend it to the semi-balanced form, which alleviates the problems including entropic bias, distance destruction, and amount constraints. We then model the measure matching under the multidimensional space, in order to learn the counting from both location and scale. To achieve this, we extend the traditional 2-D coordinate support to 3-D, incorporating an additional axis to represent scale information, where a pyramid-based structure will be leveraged to learn the scale value for the predicted density. Extensive experiments on four challenging crowd-counting datasets, namely, ShanghaiTech A, UCF-QNRF, JHU ++ , and NWPU have validated the proposed method. Code is released at https://github.com/LoraLinH/Multidimensional-Measure-Matching-for-Crowd-Counting.

Significant influence of urban human activities and marine input on rainwater chemistry in a coastal large city, China.

The coastal urban region is generally considered an atmospheric receptor for terrestrial and marine input materials, and rainfall chemistry can trace the wet scavenging process of these materials. Fast urbanization in China's east coastal areas has greatly altered the rainwater chemistry. However, the chemical variations, determinants, and sources of rainfall are unclear. Therefore, the typical coastal city of Fuzhou was selected for 1-year rainwater sampling and inorganic ions were detected to explore above problems. The findings depicted that rainwater ions in Fuzhou were slightly different from those in other coastal cities. Although NO3-, SO42-, Ca2+ and NH4+ dominated the rainwater ions, the marine input Cl- (22 %) and Na+ (11 %) also contributed a considerable percentage to the rainwater ions. Large differences in ion concentrations (2∼28 times) were found in monthly scale due to the rainfall amount. Both natural and anthropogenic determinants influenced the rainwater ions in coastal cities, such as SO2 emission, air SO2 and PM10 content on rainwater SO42-, NO3-, and Ca2+, and soot & dust emission on rainwater SO42-, NO3-, indicating the vital contribution of human activities. Stoichiometry and positive matrix factorization-based sources identification indicated that atmospheric dust/particles were the primary contributor of Ca2+ (83.3 %) and F- (83.7 %), and considerable contributor of SO42- (39.5 %), NO3- (38.3 %) and K+ (41.5 %). Anthropogenic origins, such as urban waste volatilization and fuel combustion emission, contributed 95 % of NH4+, 54.5 % of NO3- and 41.9 % of SO42-, and the traffic sources contribution was relatively higher than fixed emission sources. The marine input represented the vital source of Cl- (77.7 %), Na+ (84.9 %), and Mg2+ (55.3 %). This work highlights the significant influence of urban human activities and marine input on rainwater chemicals and provides new insight into the material cycle between the atmosphere and earth-surface in coastal city.

Thermodynamically and Kinetically Enhanced Benzene Vapor Sensor Based on the Cu-TCPP-Cu MOF with Extremely Low Limit of Detection.

As a carcinogenic and highly neurotoxic hazardous gas, benzene vapor is particularly difficult to be distinguished in BTEX (benzene, toluene, ethylbenzene, xylene) atmosphere and be detected in low concentrations due to its chemical inertness. Herein, we develop a depth-related pore structure in Cu-TCPP-Cu to thermodynamically and kinetically enhance the adsorption of benzene vapor and realize the detection of ultralow-temperature benzene gas. We find that the in-plane π electronic nature and proper pore sizes in Cu-TCPP-Cu can selectively induce the adsorption and diffusion of BTEX. Interestingly, the theoretical calculations (including density functional theory (DFT) and grand canonical Monte Carlo (GCMC) simulations) exhibit that benzene molecules are preferred to adsorb and array as a consecutive arrangement mode in the Cu-TCPP-Cu pore, while the TEX (toluene, ethylbenzene, xylene) dominate the jumping arrangement model. The differences in distribution behaviors can allow adsorption and diffusion of more benzene molecules within limited room. Furthermore, the optimal pore-depth range (60-65 nm) of Cu-TCPP-Cu allows more exposure of active sites and hinders the gas-blocking process. The optimized sensor exhibits ultrahigh sensitivity to benzene vapor (155 Hz/µg@1 ppm), fast response time (less than 10 s), extremely low limit of detection (65 ppb), and excellent selectivity (83%). Our research thus provides a fundamental understanding to design and optimize two-dimensional metal-organic framework (MOF)-based gas sensors.

Dataset Condensation via Expert Subspace Projection.

The rapid growth in dataset sizes in modern deep learning has significantly increased data storage costs. Furthermore, the training and time costs for deep neural networks are generally proportional to the dataset size. Therefore, reducing the dataset size while maintaining model performance is an urgent research problem that needs to be addressed. Dataset condensation is a technique that aims to distill the original dataset into a much smaller synthetic dataset while maintaining downstream training performance on any agnostic neural network. Previous work has demonstrated that matching the training trajectory between the synthetic dataset and the original dataset is more effective than matching the instantaneous gradient, as it incorporates long-range information. Despite the effectiveness of trajectory matching, it suffers from complex gradient unrolling across iterations, which leads to significant memory and computation overhead. To address this issue, this paper proposes a novel approach called Expert Subspace Projection (ESP), which leverages long-range information while avoiding gradient unrolling. Instead of strictly enforcing the synthetic dataset's training trajectory to mimic that of the real dataset, ESP only constrains it to lie within the subspace spanned by the training trajectory of the real dataset. The memory-saving advantage offered by our method facilitates unbiased training on the complete set of synthetic images and seamless integration with other dataset condensation techniques. Through extensive experiments, we have demonstrated the effectiveness of our approach. Our method outperforms the trajectory matching method on CIFAR10 by 16.7% in the setting of 1 Image/Class, surpassing the previous state-of-the-art method by 3.2%.

Semi-Supervised Crowd Counting via Multiple Representation Learning.

There has been a growing interest in counting crowds through computer vision and machine learning techniques in recent years. Despite that significant progress has been made, most existing methods heavily rely on fully-supervised learning and require a lot of labeled data. To alleviate the reliance, we focus on the semi-supervised learning paradigm. Usually, crowd counting is converted to a density estimation problem. The model is trained to predict a density map and obtains the total count by accumulating densities over all the locations. In particular, we find that there could be multiple density map representations for a given image in a way that they differ in probability distribution forms but reach a consensus on their total counts. Therefore, we propose multiple representation learning to train several models. Each model focuses on a specific density representation and utilizes the count consistency between models to supervise unlabeled data. To bypass the explicit density regression problem, which makes a strong parametric assumption on the underlying density distribution, we propose an implicit density representation method based on the kernel mean embedding. Extensive experiments demonstrate that our approach outperforms state-of-the-art semi-supervised methods significantly.

Microgravimetric Modeling-A New Method for Extracting Adsorption Parameters of Functionalized MIL-101(Cr).

As a volatile air pollutant, formaldehyde can enter people's living environment through interior decoration, furniture and paint, causing serious harm to human health. Therefore, it is necessary to develop a sensor for the real-time detection of formaldehyde in low concentrations. According to the chemical interaction between amino groups and formaldehyde, a MIL-101(Cr) aminated-material-based formaldehyde cantilever sensor was prepared, of which ethylenediamine- functionalized MIL-101(Cr) named ED-MIL-101(Cr)) showed the best gas sensing performance. Using quasi-in situ infrared spectroscopy, ED-MIL-101(Cr) was found bound to formaldehyde through a Schiff base. The adsorption enthalpy of formaldehyde-bound ED-MIL-101(Cr) was -52.6 kJ/mol, which corresponds to weak chemical adsorption, so the material showed good selectivity. In addition, ED-MIL-101(Cr) has the most active sites, so its response value to formaldehyde is larger and it takes longer to reach saturation adsorption than bare MIL-101(Cr). Through the research on the gas sensing performance of functionalized MIL-101(Cr) material, we found that it has a strong application potential in the field of formaldehyde monitoring, and the material performance can be quantitatively and accurately evaluated through combining calculation and experimentation for understanding the gas sensing mechanism.

Mesoporous-Structure MOF-14-Based QCM p-Xylene Gas Sensor.

In this work, a facile synthesis method was adopted to synthesize MOF-14 with mesoporous structure. The physical properties of the samples were characterized by PXRD, FESEM, TEM and FT-IR spectrometry. By coating the mesoporous-structure MOF-14 on the surface of a quartz crystal microbalance (QCM), the fabricated gravimetric sensor exhibits high sensitivity to p-toluene vapor even at trace levels. Additionally, the limit of detection (LOD) of the sensor obtained experimentally is lower than 100 ppb, and the theoretical detection limit is 57 ppb. Furthermore, good gas selectivity and fast response (15 s) and recovery (20 s) abilities are also illustrated along with high sensitivity. These sensing data indicate the excellent performance of the fabricated mesoporous-structure MOF-14-based p-xylene QCM sensor. On the basis of temperature-varying experiments, an adsorption enthalpy of -59.88 kJ/mol was obtained, implying the existence of moderate and reversible chemisorption between MOF-14 and p-xylene molecules. This is the crucial factor that endows MOF-14 with exceptional p-xylene-sensing abilities. This work has proved that MOF materials such as MOF-14 are promising in gravimetric-type gas-sensing applications and worthy of future study.

Greenhouse gas emissions and environmental drivers in different natural wetland regions of China.

Wetlands sequestrate carbon at the highest rate than any other ecosystems on Earth. However, the spatial and temporal dynamics of GHGs emissions from the wetland ecosystems in China are still elusive. We synthesized 166 publications that contain 462 in situ measurements of GHGs emissions from the natural wetlands in China, and further analyzed the variability and the drivers of GHGs emissions in eight subdivisions of China's wetlands. The results show that the current studies are mainly concentrated in the estuaries, Sanjiang Plain, and Zoige wetlands. The average CO2 emissions, CH4 fluxes and N2O fluxes from Chinese wetlands were 218.84 mg·m-2·h-1, 1.95 mg·m-2·h-1 and 5.8 × 10-2 mg·m-2·h-1, respectively. The global warming potential (GWP) of China's wetlands was estimated to be 1881.36 TgCO2-eq·yr-1, with CO2 emissions contributing more than 65% to the GWP value. The combined GWP values of Qinghai-Tibet Plateau wetlands, coastal wetlands and northeastern wetlands account for 84.8% of GWP of China's wetlands. Correlation analysis showed that CO2 emissions increased with the increasing mean annual temperature, elevation, annual rainfall, and wetland water level, but decreased with soil pH. CH4 fluxes increased with the mean annual temperature and soil water content but decreased with the redox potential. This study analyzed the drivers of GHGs emissions from wetland ecosystems at the national scale, and GWP values of eight wetland subregions of China were comprehensively assessed. Our results are potentially useful for the global GHGs inventory, and can help assess the response of GHGs emissions of wetland ecosystem to environmental and climate change.

Comparative proteomic analysis reveals differential protein expression of Hypsizygus marmoreus in response to different light qualities.

Light is important environmental stress that influences the growth, development, and metabolism of Hypsizygus marmoreus (white var.). However, the molecular basis of the light effect on H. marmoreus remains unclear. In this study, a label-free comparative proteomic analysis was applied to investigate the global protein expression profile of H. marmoreus mycelia growing under white, red, green, and blue light qualities and darkness (control). Among 3149 identified proteins in H. marmoreus, 2288 were found to be expressed in all tested conditions. Data of Each light quality was compared with darkness for further analysis, numerous differentially expressed proteins (DEPs) were identified and the white light group showed the most. All the up-regulated and down-regulated DEPs were annotated and analyzed with the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The KEGG enrichment analysis revealed that light stress was associated with primary metabolism, glycolysis/gluconeogenesis, MAPK, proteasome, and carbohydrate-active enzyme pathways. This study advances valuable insights into the molecular mechanisms underlying the role of different light qualities in mushroom growth and development.

New Application of Quartz Crystal Microbalance: A Minimalist Strategy to Extract Adsorption Enthalpy.

The capture and separation of CO2 is an important means to solve the problem of global warming. MOFs (metal-organic frameworks) are considered ideal candidates for capturing CO2, where the adsorption enthalpy is a crucial indicator for the screening of materials. For this purpose, we propose a new minimalist solution using QCM (quartz crystal microbalance) to extract the CO2 adsorption enthalpy on MOFs. Three kinds of MOFs with different properties, sizes and morphologies were employed to study the adsorption enthalpy of CO2 using a QCM platform and a commercial gas sorption analyzer. A Gaussian simulation calculation and previously data reported were used for comparison. It was found that the measuring errors were between 5.4% and 6.8%, proving the reliability and versatility of our new method. This low-cost, easy-to-use, and high-accuracy method will provide a rapid screening solution for CO2 adsorption materials, and it has potential in the evaluation of the adsorption of other gases.

Pro-angiogenic New Chloro-Azaphilone Derivatives From the Hadal Trench-Derived Fungus Chaetomium globosum YP-106.

Five new chloro-azaphilones, chaetofanixins A-E (1-5), and five known analogs (6-10) were isolated and identified from the hadal trench-derived fungus Chaetomium globosum YP-106. The structure of chaetofanixin E (5) is unique and interesting, bearing a highly rigid 6/6/5/3/5 penta-cyclic ring system, which is first encountered in natural products. The structures of these compounds, including absolute configurations, were determined based on the spectroscopic analysis, electronic circular dichroism (ECD) calculations, and analysis of biogenetic origins. Compounds 1-7 significantly promoted angiogenesis in a dose-dependent manner, and thus, these compounds might be used as promising molecules for the development of natural cardiovascular disease agents.

Sulfur-Containing Compounds from Endophytic Fungi: Sources, Structures and Bioactivities.

Endophytic fungi have attracted increasing attention as an under-explored source for the discovery and development of structurally and functionally diverse secondary metabolites. These microorganisms colonize their hosts, primarily plants, and demonstrate diverse ecological distribution. Among endophytic fungal natural products, sulfur-containing compounds feature one or more sulfur atoms and possess a range of bioactivities, e.g., cytotoxicity and antimicrobial activities. These natural products mainly belong to the classes of polyketides, nonribosomal peptides, terpenoids, and hybrids. Here, we reviewed the fungal producers, plant sources, chemical structures, and bioactivities of 143 new sulfur-containing compounds that were reported from 1985 to March 2022.

Peptide 17 alleviates early hypertensive renal injury by regulating the Hippo/YAP signalling pathway.

AIM: Hypertensive nephropathy is embodied by kidney tissue fibrosis and glomerular sclerosis, as well as renal inflammation. The Hippo/YAP (yes-associated protein, YAP) axis has been reported to promote inflammation and fibrosis and may participate in the pathogenesis of heart, vascular and renal injuries. However, the role of the Hippo/YAP pathway in hypertensive renal injury has not been reported so far. We explored the role of the Hippo/YAP signalling pathway in hypertensive renal injury and the effect of peptide 17 on its effects. METHODS: Histopathological analyses were performed based on the Masson and Haematoxylin/eosin (HE) staining approaches. Biochemical indexes were determined and immunofluorescence and western blotting were used to detect protein expression levels. The mRNA expression levels were determined by qRT-PCR. RESULTS: Our results showed that peptide 17 reduced the systolic blood pressure (SBP) and urine protein/creatinine ratio in hypertensive rats. In addition, peptide 17 reduced the histopathological damage of kidneys in spontaneously hypertensive rats (SHRs). Moreover, peptide 17 downregulated genes in the Hippo/Yap pathway in kidney tissue of SHRs and Ang II-treated kidney cells. The expression levels of inflammatory factors TNF-α, IL-1ß and MCP-1 and the pro-fibrotic factors TGF-ß1, fibronectin, and CTGF were increased in the kidney of hypertensive rats, but reversed by peptide 17 treatment. Silencing of YAP had effect similar to that of peptide 17 in vivo and in vitro. CONCLUSION: Peptide 17 alleviates early renal injury in hypertension by regulating the Hippo/YAP signalling pathway. These findings may be useful in the treatment of hypertensive renal injury.

Controllable preparation of ultrathin MXene nanosheets and their excellent QCM humidity sensing properties enhanced by fluoride doping.

A 2D ultrathin MXene nanosheet was prepared under controlled conditions and employed as a sensitive film to construct a QCM (quartz crystal microbalance) humidity sensor by a dip coating method. The MXene nanosheets were obtained by dislodging the element A from the MAX phase by a facile liquid phase etching method. The morphology and composition of the MXene nanosheets were characterized by means of a number of advanced instruments. It was found that the sample is an ultrathin graphene-like nanosheet. The sensing test results showed that the sensor has a 12.8 Hz/% RH sensitivity, 6 s and 2 s (@ 90%) response/recovery time, maximum humidity hysteresis of 1.16% RH, good stability, and selectivity. Finally, the enhanced humidity response mechanism of the MXene nanosheets was explored by density-functional theory (DFT) calculation and experimental verification. The DFT simulation together with comparison of fluoride-free sample revealed that F elements on the surface of the MXene nanosheets play a more important role in improving humidity responses than OH groups. The results present a new strategy to enhance humidity sensing performance of sensing materials by F- doping or decoration. Thus, the sensor has bright potential for humidity sensing.

Highly sensitive ethanol gas sensor based on ultrathin nanosheets assembled Bi2WO6 with composite phase.

Bismuth tungstate (Bi2WO6) has many intriguing properties and has been the focus of studies in a variety of fields, especially photocatalysis. However, its application in gas-sensing has been seldom reported. Here, we successfully synthesized assembled hierarchical Bi2WO6 which consists of ultrathin nanosheets with crystalline-amorphous composite phase by a one-step hydrothermal method. X-ray diffraction (XRD), X-ray photoemission spectroscopy (XPS), field-emission scanning electron microscopy (FESEM), and high-resolution transmission electron microscopy (HRTEM) techniques were employed to characterize its composition, morphology, and microstructure. By taking advantage of its unique microstructure, phase composition, and large surface area, we show that the resulting Bi2WO6 is capable of detecting ethanol gas with quick response (7 s) and recovery dynamic (14 s), extremely high sensitivity (Ra/Rg = 60.8@50 ppm ethanol) and selectivity. Additionally, it has excellent reproducibility and long-term stability (more than 50 d). The Bi2WO6 outperform the existing Bi2WO6-based and most of the other state-of-the-art sensing platforms. We not only provided one new member to the field of gas sensor, but also offered several strategies to reconstruct nanomaterials.

Cigarette smoking and chronic kidney disease in the general population: a systematic review and meta-analysis of prospective cohort studies.

Background: Cigarette smoking has been established as a major risk factor for chronic kidney disease (CKD) development in people with diabetes. Conflicting evidence exists among representative community-based studies as to whether smoking is an independent risk factor for CKD. The aim of this meta-analysis was to assess the effects of tobacco smoking on the development of CKD in adult general populations. Methods: A literature search was conducted using MEDLINE and Embase from their inception through 31 May 2016 for prospective cohort studies that reported relative risks of CKD with smoking status in the general population. Summary relative risks (SRRs) and 95% confidence intervals (CIs) were calculated using a random effects model. Results: A total of 15 prospective cohort studies, including 65 064 incident CKD cases, were included. Compared with never-smokers, the SRRs of incident CKD were 1.27 (95% CI 1.19-1.35) for ever-smokers, 1.34 (95% CI 1.23-1.47) for current smokers and 1.15 (95% CI 1.08-1.23) for former smokers. The SRRs for end-stage renal disease development were 1.51 (95% CI 1.24-1.84) for ever-smokers, 1.44 (95% CI 1.00-2.09) for former smokers and 1.91 (95% CI 1.39-2.64) for current smokers. Considerable heterogeneity was observed among these studies. After combining an additional three prospective cohort studies, which involved 5747 incident proteinuria cases, a null smoking-proteinuria association in the general population was observed. Conclusions: Our meta-analysis suggests evidence for cigarette smoking as an independent risk factor for incident CKD. Future studies are required to investigate whether smoking cessation can decrease incident CKD in the general adult population.

[Effect of Kangxianling Recipe on p38MAPK/NF-KBp65 Mediated Inflammatory Factors in 5/6 Nephrectomized Mice].

Objective To observe effects of Kangxianling Recipe ( KXLR) on p38MAPK/NF- κBp65 mediated inflammatory factors in chronic renal failure ( CRF) model mice. Methods Totally 56 C57BL/6J male mice (18 -22 g) were recruited in this experiment. Ten were randomly selected as a sham-operation group. The rest 46 mice were used for preparing CRF model by 5/6 nephrectomization. To- tally 33 successfully modeled mice were divided into the model group, the rapamycin (RAP) group, and the KXLR group according to serum creatinine (SCr) level, 11 in each group. Mice in the RAP group were administered with rapamycin (0.13 mg/100 g per day, 0. 5 mL each time) by gastrogavage. Mice in the KXLR group were administered with KXLR (2 g/100 g per day, 0. 5 mL each time) by gastrogavage. Equal volume of distilled water was administered to mice in the model group and the sham-operation group. Mice were sacrificed after 8 weeks of consecutive medication. The expression of neutrophils was ob- served using immunohistochemical assay. Expression levels of p38MAPK/NF-κB p65 protein and TNF-α/ IL-6 mRNA were detected by Western blot and Real-time PCR. Results Compared with the sham-opera- tion group, the number of positive neutrophils increased, expression levels of p38MAPK/NF-κB p65 protein and TNF-α/IL-6 mRNA were enhanced significantly in the model group (P <0. 05, P <0. 01). Com- pared with the model group, the number of neutrophils was reduced, expression levels of p38MAPK/NF- κB p65 protein and TNF-α/IL-6 mRNA were decreased significantly in the KXLR group and the RAP group (P <0. 05, P <0. 01). RAP showed better effect in decreasing p38MAPK protein expression than KXLR (P <0. 05). There was no statistical difference in the rest indices between the KXLR group and the RAP group (P >0. 05). Conclusions KXLR participated the regulation of p38MAPK/NF-κB p65 mediated in- flammation factors. It had certain improvement in renal fibrosis induced renal failure.

CXCL16 regulates renal injury and fibrosis in experimental renal artery stenosis.

Recent studies have shown that inflammation plays a critical role in the initiation and progression of hypertensive kidney disease, including renal artery stenosis. However, the signaling mechanisms underlying the induction of inflammation are poorly understood. We found that CXCL16 was induced in the kidney in a murine model of renal artery stenosis. To determine whether CXCL16 is involved in renal injury and fibrosis, wild-type and CXCL16 knockout mice were subjected to renal artery stenosis induced by placing a cuff on the left renal artery. Wild-type and CXCL16 knockout mice had comparable blood pressure at baseline. Renal artery stenosis caused an increase in blood pressure that was similar between wild-type and CXCL16 knockout mice. CXCL16 knockout mice were protected from RAS-induced renal injury and fibrosis. CXCL16 deficiency suppressed bone marrow-derived fibroblast accumulation and myofibroblast formation in the stenotic kidneys, which was associated with less expression of extracellular matrix proteins. Furthermore, CXCL16 deficiency inhibited infiltration of F4/80(+) macrophages and CD3(+) T cells in the stenotic kidneys compared with those of wild-type mice. Taken together, our results indicate that CXCL16 plays a pivotal role in the pathogenesis of renal artery stenosis-induced renal injury and fibrosis through regulation of bone marrow-derived fibroblast accumulation and macrophage and T-cell infiltration.

CXCL16 Deficiency Attenuates Renal Injury and Fibrosis in Salt-Sensitive Hypertension.

Inflammation plays an important role in the pathogenesis of hypertensive kidney disease. However, the molecular mechanisms underlying the induction of inflammation are not completely understood. We have found that CXCL16 is induced in the kidney in deoxycorticosterone acetate (DOCA)-salt hypertension. Here we examined whether CXCL16 is involved in DOCA-salt-induced renal inflammation and fibrosis. Wild-type and CXCL16 knockout mice were subjected to uninephrectomy and DOCA-salt treatment for 3 weeks. There was no difference in blood pressure at baseline between wild-type and CXCL16 knockout mice. DOCA-salt treatment resulted in significant elevation in blood pressure that was comparable between wild-type and CXCL16 knockout mice. CXCL16 knockout mice exhibited less severe renal dysfunction, proteinuria, and fibrosis after DOCA-salt treatment compared with wild-type mice. CXCL16 deficiency attenuated extracellular matrix protein production and suppressed bone marrow-derived fibroblast accumulation and myofibroblast formation in the kidneys following DOCA-salt treatment. Furthermore, CXCL16 deficiency reduced macrophage and T cell infiltration into the kidneys in response to DOCA-salt hypertension. Taken together, our results indicate that CXCL16 plays a key role in the pathogenesis of renal injury and fibrosis in salt-sensitive hypertension through regulation of bone marrow-derived fibroblast accumulation and macrophage and T cell infiltration.