Genome-wide investigation to assess copy number variants in the Italian local chicken population

Background Copy number variants(CNV)hold significant functional and evolutionary importance.Numerous ongoing CNV studies aim to elucidate the etiology of human diseases and gain insights into the population structure of livestock.High-density chips have enabled the detection of CNV with increased resolution,leading to the identification of even small CNV.This study aimed to identify CNV in local Italian chicken breeds and investigate their distribution across the genome.Results Copy number variants were mainly distributed across the first six chromosomes and primarily associated with loss type CNV.The majority of CNV in the investigated breeds were of types 0 and 1,and the minimum length of CNV was significantly larger than that reported in previous studies.Interestingly,a high proportion of the length of chromosome 16 was covered by copy number variation regions(CNVR),with the major histocompatibility complex being the likely cause.Among the genes identified within CNVR,only those present in at least five animals across breeds(n=95)were discussed to reduce the focus on redundant CNV.Some of these genes have been associated to functional traits in chickens.Notably,several CNVR on different chromosomes harbor genes related to muscle development,tissue-specific biological processes,heat stress resistance,and immune response.Quantitative trait loci(QTL)were also analyzed to investigate potential overlapping with the identified CNVR:54 out of the 95 gene-containing regions overlapped with 428 QTL associated to body weight and size,carcass characteristics,egg production,egg components,fat deposition,and feed intake.Conclusions The genomic phenomena reported in this study that can cause changes in the distribution of CNV within the genome over time and the comparison of these differences in CNVR of the local chicken breeds could help in preserving these genetic resources.

Genetic variants in C1GALT1 are associated with gastric cancer risk by influencing immune infiltration

Core 1 synthase glycoprotein-N-acetylgalactosamine 3-β-galactosyltransferase 1(C1GALT1)is known to play a critical role in the development of gastric cancer,but few studies have elucidated associations between genetic variants in C1GALT1 and gastric cancer risk.By using the genome-wide association study data from the database of Genotype and Phenotype(dbGAP),we evaluated such associations with a multivariable logistic regression model and identified that the rs35999583 G>C in C1GALT1 was associated with gastric cancer risk(odds ratio,0.83;95% confidence interval[CI],0.75-0.92;P=3.95×10^(-4)).C1GALT1 mRNA expression levels were significantly higher in gastric tumor tissues than in normal tissues,and gastric cancer patients with higher C1GALT1 mRNA levels had worse overall survival rates(hazards ratio,1.33;95%CI,1.05-1.68;P_(log-rank)=1.90×10^(-2)).Furthermore,we found that C1GALT1 copy number differed in various immune cells and that C1GALT1 mRNA expression levels were positively correlated with the infiltrating levels of CD4^(+)T cells and macrophages.These results suggest that genetic variants of C1GALT1 may play an important role in gastric cancer risk and provide a new insight for C1GALT1 into a promising predictor of gastric cancer susceptibility and immune status.

Contribution of Genomic Surveillance in the Detection and Monitoring of SARS-CoV-2 Variants during the 6 Pandemic Waves in the Central African Republic from 2020 to 2023

Objective: The COVID-19 pandemic has highlighted the need to strengthen diagnosis and genomic surveillance capacities. In 2021, Central African managed five waves of COVID-19 by integrating genomic surveillance into their health monitoring system. This study sought to report surveillance data from the National Laboratory of Clinical Biology and Public Health and describe the circulation of SARS-CoV-2 variants. Materials and Methods: This retrospective, descriptive observational study spans three years, from April 2020 to November 2023. It was conducted on a population of consenting volunteers from across the Central African Republic, who were tested using RT-PCR on nasopharyngeal samples. Data with sufficient information were obtained from the National Laboratory of Clinical Biology and Public Health (LNBCSP) databases. Sequencing was largely carried out at the National Institute of Biomedical Research (INRB) in Kinshasa until May 2023, and subsequently at the LNBCSP. Results and Discussion: Out of 97,864 RT-PCR tests performed, 9,764 were positive, resulting in a prevalence of 9.98%. The average age of the patients was 39.97 years ± 13.76, and the male-to-female sex ratio was 2.12. RT-PCR test positivity was significantly associated with age (p = 0.001), sex (p = 0.013) and clinical manifestations. Ten variants circulated during the five recorded waves, with Omicron (B.1.1.529), Delta (B.1.617.2) variants being predominant. Notably, the B.1.620 and B.640 variants were prominent during the second wave. Conclusion: This retrospective study provides key insights into the COVID-19 pandemic in the CAR. It identifies risk factors and details the circulation of various SARS-CoV-2 variants. Enhancing national genomic surveillance capacities would enable the country to better respond to future pandemic challenges.

New recessive compound heterozygous variants of RP1L1 in RP1L1 maculopathy

AIM:To identify a maculopathy patient caused by new recessive compound heterozygous variants in RP1L1.METHODS:Comprehensive retinal morphological and functional examinations were evaluated for the patient with RP1L1 maculopathy.Targeted sequence capture array technique was used to screen potential pathologic variants.Polymerase chain reaction and Sanger sequencing were used to confirm the screening results.RESULTS:Fundus examination showed round macular lesions appeared in both eyes.Optical coherence tomography showed that the inner segment/outer segment continuity was disorganized and disruptive in the left eye,but it was uneven and slightly elevated in the right eye.Fundus autofluorescence showed patchy hyper-autofluorescence in the macula.Visual field examination indicates central defects in both eyes.Electroretinogram(ERG)and multifocal ERG showed no obvious abnormalities.Fundus fluorescein angiography in the macula showed obviously irregular hyper-fluorescence in the right eye and slightly hyper-fluorescence in the left eye.We found that the proband carried a missense variant(c.1972C>T)and a deletion variant(c.4717_4718del)of RP1L1,which were originated from the parents and formed compound heterozygous variants.Both variants are likely pathogenic according to the ACMG criteria.Multimodal imaging,ERG and detailed medical history are important diagnostic tools for differentiating between acquired and inherited retinal disorders.CONCLUSION:A maculopathy case with detailed retinal phenotype and new recessive compound heterozygous variants of RP1L1 is identified in a Chinese family,which expands the understanding of phenotype and genotype in RP1L1 maculopathy.

Sex Determination in Homo sapiens as a Multi-Step Process: Potential for Development of Variants and Sex Differences in Disease Risk

Reproduction via cis-binary mechanisms appears to have evolved fairly early in the evolution of complex organisms, and a system committed to prior to evolution of humans. While the evolution of a chromosomal-specific approach has been a successful strategy for survival of a large variety of species including humans, the fidelity of sex determination leading to 100% cis-binary outcomes is not achieved in many species, with evidence for homosexual or bisexual behaviour evident in more than 1500 species. Thus, such outcomes indicates that sex determination is a multi-step process and not a single event, and as such, could lead to the appearance of variants during the process which developed much earlier than humans. Variants could arise either due to intrinsic variation in the steps of determination, or also be influenced by environmental factors of a biological or psychological nature. In contrast to homosexual variants which do not require interventions such as hormone therapy or surgery, expression of gender dysphoria, is more based in psychology, but also has biological underpinnings and can be influenced by such hormonal interventions and surgery. While the numbers of those with gender dysphoria is small (~0.6% - 1.0% of population), the attention given to this issue raises the possibility of biological and psychological environmental factors impacting the emergence of some of those expressing gender dysphoria. Furthermore, transitioning from male-to-female or female-to-male can have consequences regarding disease risks latter in life, including the appearance of autoimmune diseases. This review will attempt to review some of the evidence regarding sex determination, discuss why the system has potentially not been improved upon during evolution, how a potential role for sex chromosome function on neurodevelopment may be central to variation in humans, and how commitment to the current strategy is likely integrated into other sex-related events such as puberty, pregnancy, and menopause to ensure species survival. It will also discuss how variants in sex determination could contribute to sex differences in disease risk and how epigenetic modifications could play a role in such risk. .

Lab results of COVID-19 patients:Omicron vs delta variants

BACKGROUND The coronavirus disease 2019(COVID-19)virus has been a world-known pan-demic since February 2020.Multiple variances had been established;the most common variants in Israel were omicron and delta.AIM To analyze and compare laboratory values in the"omicron"and"delta"variants of the coronavirus by conducting follow-up examinations and laboratory audits on COVID-19 patients admitted to our institution.METHODS A retrospective study,two groups,50 patients in each group.Patients examined positive for COVID-19 were divided into groups according to the common variant at the given time.We reviewed demographic data and laboratory results such as complete blood count and full chemistry,including electrolytes and coagulation parameters.RESULTS The mean age was 52%,66.53±21.7 were female.No significance was found comparing laboratory results in the following disciplines:Blood count,hemo-globin,and lymphocytes(P=0.41,P=0.87,P=0.97).Omicron and delta variants have higher neutrophil counts,though they are not significantly different(P=0.38).Coagulation tests:Activated paritial thromoplastin test and international normalized ratio(P=0.72,P=0.68).We found no significance of abnormality for all electrolytes.CONCLUSION The study compares laboratory results of blood tests between two variants of the COVID-19 virus–omicron and delta.We found no significance between the variants.Our results show the need for further research with larger data as well as the need to compare all COVID-19 variants.

Genetic Analysis of Two Novel GPI Variants Disrupting H Bonds and Localization Characteristics of 55 Gene Variants Associated with Glucose-6-phosphate Isomerase Deficiency

Objective:Glucose-6-phosphate isomerase(GPI)deficiency is a rare hereditary nonspherocytic hemolytic anemia caused by GPI gene variants.This disorder exhibits wide heterogeneity in its clinical manifestations and molecular characteristics,often posing challenges for precise diagnoses using conventional methods.To this end,this study aimed to identify the novel variants responsible for GPI deficiency in a Chinese family.Methods:The clinical manifestations of the patient were summarized and analyzed for GPI deficiency phenotype diagnosis.Novel compound heterozygous variants of the GPI gene,c.174C>A(p.Asn58Lys)and c.1538G>T(p.Trp513Leu),were identified using whole-exome and Sanger sequencing.The AlphaFold program and Chimera software were used to analyze the effects of compound heterozygous variants on GPI structure.Results:By characterizing 53 GPI missense/nonsense variants from previous literature and two novel missense variants identified in this study,we found that most variants were located in exons 3,4,12,and 18,with a few localized in exons 8,9,and 14.This study identified novel compound heterozygous variants associated with GPI deficiency.These pathogenic variants disrupt hydrogen bonds formed by highly conserved GPI amino acids.Conclusion:Early family-based sequencing analyses,especially for patients with congenital anemia,can help increase diagnostic accuracy for GPI deficiency,improve child healthcare,and enable genetic counseling.

Functional analysis of the novel mitochondrial tRNA^(Trp)and tRNA^(Ser(AGY))variants associated with type 2 diabetes mellitus

BACKGROUND Mutations in mitochondrial tRNA(mt-tRNA)genes that result in mitochondrial dysfunction play important roles in type 2 diabetes mellitus(T2DM).We previously reported a large Chinese pedigree with maternally inherited T2DM that harbors novel mt-tRNA^(Trp)A5514G and tRNA^(Ser(AGY))C12237T variants,however,the effects of these mt-tRNA variants on T2DM progression are largely unknown.AIM To assess the potential pathogenicity of T2DM-associated m.A5514G and m.C12237T variants at genetic,molecular,and biochemical levels.METHODS Cytoplasmic hybrid(cybrid)cells carrying both m.A5514G and m.C12237T variants,and healthy control cells without these mitochondrial DNA(mtDNA)variants were generated using trans-mitochondrial technology.Mitochondrial features,including mt-tRNA steady-state level,levels of adenosine triphosphate(ATP),mitochondrial membrane potential(MMP),reactive oxygen species(ROS),mtDNA copy number,nicotinamide adenine dinucleotide(NAD+)/NADH ratio,enzymatic activities of respiratory chain complexes(RCCs),8-hydroxy-deoxyguanine(8-OhdG),malondialdehyde(MDA),and superoxide dismutase(SOD)were examined in cell lines with and without these mt-tRNA variants.RESULTS Compared with control cells,the m.A5514G variant caused an approximately 35%reduction in the steady-state level of mt-tRNA^(Trp)(P<0.0001);however,the m.C12237T variant did not affect the mt-tRNA^(Ser(AGY))steady-state level(P=0.5849).Biochemical analysis revealed that cells with both m.A5514G and m.C12237T variants exhibited more severe mitochondrial dysfunctions and elevated oxidative stress than control cells:ATP,MMP,NAD+/NADH ratio,enzyme activities of RCCs and SOD levels were markedly decreased in mutant cells(P<0.05 for all measures).By contrast,the levels of ROS,8-OhdG and MDA were significantly increased(P<0.05 for all measures),but mtDNA copy number was not affected by m.A5514G and m.C12237T variants(P=0.5942).CONCLUSION The m.A5514G variant impaired mt-tRNA^(Trp)metabolism,which subsequently caused mitochondrial dysfunction.The m.C12237T variant did not alter the steady-state level of mt-tRNA^(Ser(AGY)),indicating that it may be a modifier of the m.A5514G variant.The m.A5514G variant may exacerbate the pathogenesis and progression of T2DM in this Chinese pedigree.

Integration between Genomic and Computational Statistical Surveys for the Screening of SNP Genetic Variants in Inflammatory Bowel Disease (IBD) Pediatric Patients*

Inflammatory bowel diseases (IBD) are complex multifactorial disorders that include Crohn’s disease (CD) and ulcerative colitis (UC). Considering that IBD is a genetic and multifactorial disease, we screened for the distribution dynamism of IBD pathogenic genetic variants (single nucleotide polymorphisms;SNPs) and risk factors in four (4) IBD pediatric patients, by integrating both clinical exome sequencing and computational statistical approaches, aiming to categorize IBD patients in CD and UC phenotype. To this end, we first aligned genomic read sequences of these IBD patients to hg19 human genome by using bowtie 2 package. Next, we performed genetic variant calling analysis in terms of single nucleotide polymorphism (SNP) for genes covered by at least 20 read genomic sequences. Finally, we checked for biological and genomic functions of genes exhibiting statistically significant genetic variant (SNPs) by introducing Fitcon genomic parameter. Findings showed Fitcon parameter as normalizing IBD patient’s population variability, as well as inducing a relative good clustering between IBD patients in terms of CD and UC phenotypes. Genomic analysis revealed a random distribution of risk factors and as well pathogenic SNPs genetic variants in the four IBD patient’s genome, claiming to be involved in: i) Metabolic disorders, ii) Autoimmune deficiencies;iii) Crohn’s disease pathways. Integration of genomic and computational statistical analysis supported a relative genetic variability regarding IBD patient population by processing IBD pathogenic SNP genetic variants as opposite to IBD risk factor variants. Interestingly, findings clearly allowed categorizing IBD patients in CD and UC phenotypes by applying Fitcon parameter in selecting IBD pathogenic genetic variants. Considering as a whole, the study suggested the efficiency of integrating clinical exome sequencing and computational statistical tools as a right approach in discriminating IBD phenotypes as well as improving inflammatory bowel disease (IBD) molecular diagnostic process.

Clinicopathological features and medium-term outcomes of histologic variants of primary focal segmental glomerulosclerosis in adults:A retrospective study

BACKGROUND The Columbia classification identified five histological variants of focal segmental glomerulosclerosis(FSGS).The prognostic significance of these variants remains controversial.AIM To evaluate the relative frequency,clinicopathologic characteristics,and medium-term outcomes of FSGS variants at a single center in Pakistan.METHODS This retrospective study was conducted at the Department of Nephrology,Sindh Institute of Urology and Transplantation,Karachi,Pakistan on all consecutive adults(≥16 years)with biopsy-proven primary FSGS from January 1995 to December 2017.Studied subjects were treated with steroids as a first-line therapy.The response rates,doubling of serum creatinine,and kidney failure(KF)with replacement therapy were compared between histological variants using ANOVA or Kruskal Wallis,and Chi-square tests as appropriate.Data were analyzed by SPSS version 22.0.P-value≤0.05 was considered significant.RESULTS A total of 401 patients were diagnosed with primary FSGS during the study period.Among these,352(87.7%)had a designated histological variant.The not otherwise specified(NOS)variant was the commonest,being found in 185(53.9%)patients,followed by the tip variant in 100(29.1%)patients.Collapsing(COL),cellular(CEL),and perihilar(PHI)variants were seen in 58(16.9%),6(1.5%),and 3(0.7%)patients,respectively.CEL and PHI variants were excluded from further analysis due to small patient numbers.The mean follow-up period was 36.5±29.2 months.Regarding response rates of variants,patients with TIP lesions achieved remission more frequently(59.5%)than patients with NOS(41.8%)and COL(24.52%)variants(P<0.001).The hazard ratio of complete response among patients with the COL variant was 0.163[95%confidence interval(CI):0.039-0.67]as compared to patients with NOS.The TIP variant showed a hazard ratio of 2.5(95%CI:1.61-3.89)for complete remission compared to the NOS variant.Overall,progressive KF was observed more frequently in patients with the COL variant,43.4%(P<0.001).Among these,24.53%of patients required kidney replacement therapy(P<0.001).The hazard ratio of doubling of serum creatinine among patients with the COL variant was 14.57(95%CI:1.87-113.49)as compared to patients with the TIP variant.CONCLUSION In conclusion,histological variants of FSGS are predictive of response to treatment with immunosuppressants and progressive KF in adults in our setup.

基于改进U-net的少样本煤岩界面图像分割方法

煤岩图像语义分割技术是煤岩界面识别的重要研究方向,现有的语义分割模型通常依赖于大样本数据集进行训练,然而目前已标注的煤岩图像数据样本难以获取,并且缺乏公开数据集。针对以上问题,提出了一种基于改进U-net模型的样本煤岩界面图像分割模型。将裁剪后具有更强特征提取能力且结构上更为简单的VGG16替换U-net的原始骨干特征提取网络,提升对图像信息的特征提取能力并获得更快的训练速度,在U-net网络的跳跃连接和解码器上采样部分引入注意力机制模块,对提取的特征层进行处理,提升模型对煤岩界面图像关键特征的提取能力,提高分割精度。使用迁移学习方法对改进的模型进行预训练,提高模型泛化能力同时避免过拟合,使模型更适用于小样本数据集训练。通过使用自制的煤岩界面数据集对所改进的网络模型性能进行验证,并将该模型与经典Unet、DeepLabv3+、PspNet、HrNet网络模型进行了对比。试验结果表明:在同样使用由125幅煤岩界面图片构建的小样本数据集进行训练的情况下,所提改进模型相较于经典U-net模型在分割精确度和检测效率方面都有显著提升,模型精确度提高了1.84%,平均交并比提高了5.34%,类别平均像素准确率提高了0.48%,检测速度增幅为5.3%。同时,与其他网络模型相比,所提改进模型在小样本煤岩界面图像的语义分割中优势显著,表明所提改进思路的有效性。

基于改进U-Net的根系表型参数测量系统

为解决背景噪声干扰下,从微根管采集的原位根系图像中难以直接提取准确的表型参数问题,提出一种基于改进U-Net的微根管根系表型参数测量系统。在U-Net网络中引入优化后的空洞空间金字塔池化模块(Atrous Spatial Pyramid Pooling,ASPP)和高效通道注意力模块(Efficient Channel Attention,ECA),增大感受野,提升模型捕捉根系细节特征的能力,获取精确的根系分割图像。结果表明,改进的U-Net模型平均交并比和平均像素精度分别为87.07%和91.85%,相较原始U-Net分别提高了2.49%和2.3%。与WinRHIZO根系分析软件测量值相比,根长度和面积决定系数分别为0.951 8和0.984 9,Spearman相关系数分别为0.972 5和0.975 7,可以实现根系长度和面积的准确测量。

面向青花瓷碎片图像的U-Net++拼接网络

针对现有图像拼接方法存在拼接处伪影以及非重叠区域内容失真,导致较低的准确性和鲁棒性的问题,提出一种基于U-Net++消除伪影的青花瓷碎片图像拼接方法.首先估计待拼接图像单应性矩阵;然后将单应性矩阵应用于结构拼接阶段,得到图像粗拼接结果;最后以图像粗拼接结果作为先验信息,在内容校正阶段改进现有的U-Net,利用U-Net++细化粗拼接结果,得到最终图像精确拼接.以青花瓷碎片图像数据集与相关经典方法进行实验的结果表明,在3个评价指标中,所提方法的峰值信噪比提高约13%,均方根误差降低约33%,均方误差降低57%左右;该方法具有较小的误差比,不仅能够提高图像拼接质量,而且表现出较好的鲁棒性.

基于残差U-Net和自注意力Transformer编码器的磁场预测方法

利用有限元方法对几何结构复杂的电机和变压器进行磁场分析,存在仿真时间长且无法复用的问题。因此,该文提出一种基于残差U-Net和自注意力Transformer编码器的磁场预测方法。首先建立永磁同步电机(PMSM)和非晶合金变压器(AMT)有限元模型,得到深度学习训练所需的数据集;然后将Transformer模块与U-Net模型结合,并引入短残差机制建立ResUnet-Transformer模型,通过预测图像的像素实现磁场预测;最后通过Targeted Dropout算法和动态学习率调整策略对模型进行优化,解决拟合问题并提高预测精度。计算实例证明,ResUnet-Transformer模型在PMSM和AMT数据集上测试集的平均绝对百分比误差(MAPE)均小于1%,且仅需500组样本。该文提出的磁场预测方法能减少实际工况和多工况下精细模拟和拓扑优化的时间和资源消耗,亦是虚拟传感器乃至数字孪生的关键实现方法之一。

一种改进U-Net网络的心电图分类算法研究

基于CPSC-2018十二导联数据,提出了一种U-Net网络和注意力机制结合的心电图分类算法。首先,针对数据集数据长度长短不一的问题,对数据进行等长处理和归一化处理。然后,利用U-Net网络中跳层连接和编码解码方式,对预处理后较长的数据进行处理。在U-Net网络解码的最后一层加入注意力机制对抗噪声,提升模型的有效信息关注度和准确性。最后,利用CPSC-2018数据集进行验证。实验结果表明:所提模型能够取得较好的分类效果,识别房颤(AF)和右束支传导阻滞(RBBB)心律失常的精准率、召回率、F1值都可以达到90%以上,平均F1值可以达到82.5%。

基于AttentionR2U-net的岩石(体)关键节理智能识别与参数提取

针对岩石(体)表面复杂节理网中关键节理的智能识别与参数提取问题,提出一种基于AttentionR2U-net网络与节理几何特征模型耦合识别的方法.在R2U-net网络的基础上引入注意门(attentiongate)改进网络,通过定性与定量的方法对边坡节理图像和混凝土、龟裂土、常见脆性岩石裂隙图像的识别结果分别作准确性及泛化能力检验;利用AttentionR2U-net网络耦合节理几何特征的方法识别关键节理,提取原始节理和关键节理的几何参数并对其迹长、面积及倾角作差异性分析.研究结果表明:针对岩石(体)节理识别,本文算法的Dice相似系数从U-net网络的0.965提升至0.990,且明显优于传统算法,故本文算法在岩石(体)节理识别上具有更强的可靠性与优越性;针对混凝土、龟裂土和大理岩、花岗岩、砂岩等脆性岩石裂隙的识别,本文算法的Dice相似系数均在0.953以上,故本文算法具有较强的泛化能力.与原始节理网络相比,关键节理网络优势迹长由0.732m显著增大至1.835m,节理倾角分布形式和优势倾角组均不变,优势迹长和倾角的节理占比均显著增大.

基于U-net神经网络的油浸式变压器绕组流-热耦合快速计算

该文针对采用传统数值方法进行大型油浸变压器绕组温升仿真时间较长的问题,提出一种基于U-net神经网络训练的快速计算方法,可以迅速地预测变压器绕组温升及热点。首先,根据流热耦合原理筛选输入变量,并运用流热耦合方法计算不同工况下的输出结果,并将之制作成训练集和测试集。同时,详细讨论3个对网络训练影响最显著的超参数;其次,将归一化后的训练集输入U-net神经网络进行训练,并设置超参数最佳组合;最后,将预测集输入训练好的模型进行预测计算及反归一化操作,预测绕组热点与Fluent仿真结果相差仅0.44 K,单次仿真时间从200 s缩短为0.07 s。预测结果与实验温度平均误差最大为2.31 K,最小为0.98 K,预测方差为0.31左右。结果表明:该方法可用于快速获得油浸式变压器绕组的温度及热点,可满足变压器温度热点数字孪生的实时性仿真要求。

基于改进Faster R-CNN与U-Net算法的桥梁病害识别与量化方法

为实现桥梁病害检测的自动化,对基于图像处理技术的混凝土桥梁表观病害的智能识别和尺寸确定方法展开研究.提出基于改进Faster R-CNN算法的病害识别方法,利用K均值聚类和遗传算法对区域候选网络锚框进行优化设计;以裂缝预测区域为基础,提出ResNet34结合U-Net的裂缝形态提取方法,并结合裂缝形态学研究了裂缝像素宽度和长度的确定方法.结果表明:锚框优化设计可改进Faster R-CNN算法的表观病害识别效果,5类常见病害的预测准确率、召回率、平均精确率分别由68.40%、69.87%、74.64%提升到85.40%、83.59%、83.72%;利用病害预测框,结合改进U-Net算法的裂缝像素尺寸计算,可实现裂缝病害尺寸的自动测量;基于改进Faster R-CNN和改进U-Net的方法可实现混凝土桥梁常见病害的智能识别和尺寸量化,从而提高桥梁病害检测效率并促进桥梁技术状况评定的智能化.

应用自适应注意力机制U-net的地震数据高分辨处理

随着油气勘探开发的不断深入,薄储层与岩性油气藏逐渐成为重要的勘探目标,这也对地震资料的分辨率提出了更高的要求。文中提出了一种基于自适应注意力机制的U-net地震数据高分辨处理方法。该方法首先利用U-net结构学习地震数据的特征表示,通过下采样过程的编码器提取地震数据的抽象特征,然后通过上采样的解码器进行特征重建和细化。在上采样的过程中引入了注意力机制,用于自适应地调整网络对不同地震特征的关注程度,网络能够更加有效地捕捉到地震数据更多的细节和特征。Marmousi模型合成地震记录和实际数据实验结果表明,新网络比原U-net误差更小、更稳定,可有效提高预测精度,实现对地震数据的高分辨率处理。

基于MSCAU-Net的视网膜眼底图像的硬渗出液分割

硬渗出液是早期糖尿病性视网膜病变(diabetic retinopathy,DR)的主要病症之一,在眼底图像中占据的像素点较少,其检测容易受视盘、软渗出液的干扰。针对这些问题,在U型网络(U-Net)结构的基础上,通过在编码器和解码器中融入残差模块和残差通道注意力模块学习硬渗出液的细微特征,在跳跃连接中加入一种新的多尺度通道注意力(multi-scale channel attention,MSCA)模块提升网络对稀疏小病灶的分割能力,提出了MSCA U-Net。基于超广角眼底图像数据集和印度糖尿病性视网膜病变图像数据集的实验结果表明,与其他基于卷积神经网络的图像分割方法相比,所提方法具有更高的硬渗出液分割精度。