Annotated Data Research Articles

Class-aware multi-level attention learning for semi-supervised breast cancer diagnosis under imbalanced label distribution.

Breast cancer affects a significant number of patients worldwide, and early diagnosis is critical for improving cure rates and prognosis. Deep learning-based breast cancer classification algorithms have substantially alleviated the burden on medical personnel. However, existing breast cancer diagnosis models face notable limitations which are challenging to obtain in clinical settings, such as reliance on a large volume of labeled samples, an inability to comprehensively extract features from breast cancer images, and susceptibility to overfitting on account of imbalanced class distribution. Therefore, we propose the class-aware multi-level attention learning model focused on semi-supervised breast cancer diagnosis to effectively reduce the dependency on extensive data annotation. Additionally, we develop the multi-level fusion attention learning module, which integrates multiple mutual attention components across different layers, allowing the model to precisely identify critical regions for lesion categorization. Finally, we design the class-aware adaptive pseudo-labeling module which adaptively predicts category distribution in unlabeled data, and directs the model to focus on underrepresented categories, ensuring a balanced learning process. Experimental results on the BACH dataset demonstrate that our proposed model achieves an accuracy of 86.7% with only 40% labeled microscopic data, showcasing its outstanding contribution to semi-supervised breast cancer diagnosis.

A novel cross-modal data augmentation method based on contrastive unpaired translation network for kidney segmentation in ultrasound imaging.

Kidney ultrasound (US) image segmentation is one of the key steps in computer-aided diagnosis and treatment planning of kidney diseases. Recently, deep learning (DL) technology has demonstrated promising prospects in automatic kidney US segmentation. However, due to the poor quality, particularly the weak boundaries in kidney US imaging, obtaining accurate annotations for DL-based segmentation methods remain a challenging and time-consuming task. This issue can hinder the application of data-hungry deep learning methods. In this paper, we explore a novel cross-modal data augmentation method aimed at enhancing the performance of DL-based segmentation networks on the limited labeled kidney US dataset. In particular, we adopt a novel method based on contrastive unpaired translation network (CUT) to obtain simulated labeled kidney US images at a low cost from labeled abdomen computed tomography (CT) data and unlabeled kidney US images. To effectively improve the segmentation network performance, we propose an instance-weighting training strategy that simultaneously captures useful information from both the simulated and real labeled kidney US images. We trained our generative networks on a dataset comprising 4418 labeled CT slices and 4594 unlabeled US images. For segmentation network, we used a dataset consisting of 4594 simulated and 100 real kidney US images for training, 20 images for validation, and 169 real images for testing. We compared the performance of our method to several state-of-the-art approaches using the Wilcoxon signed-rank test, and applied the Bonferroni method for multiple comparison correction. The experimental results show that we can synthesize accurate labeled kidney US images with a Fréchet inception distance of 52.52. Moreover, the proposed method achieves a segmentation accuracy of 0.9360±0.0398 for U-Net on normal kidney US images, and 0.7719±0.2449 on the abnormal dataset, as measured by the dice similarity coefficient. When compared to other training strategies, the proposed method demonstrated statistically significant superiority, with all p-values being less than 0.01. The proposed method can effectively improve the accuracy and generalization ability of kidney US image segmentation models with limited annotated training data.

Aerobic Stress Detection in Aquatic Environments with Water Quality Data Using Hybrid Deep Learning Based ConvRec Model

Depletion of dissolved oxygen in the water is a serious threat to fish and other aquatic organisms, it causes aerobic stress disease in fish. Detection of aerobic stress is crucial to maintain better growth and spawning in the fishes. Recently many studies proposed deep learning-based water quality analysis techniques, but these techniques inadequate in handling the complex water quality data. Because water quality has both spatial and temporal characteristics, this makes most of the deep learning models inadequate. To handle such complex and multifaceted data we proposed ConvRec, a deep learning architecture that incorporates CNN (Convolution neural network) and LSTM (Long-short term network) structures. CNN component extracts feature in the spatial domain from the water quality data from different locations while LSTM captures temporal features hence the model can learn both spatial and temporal correlations between the movement of water quality parameters to classify the aerobic stress in aqua ponds. In this work we use the two dataset both are unlabelled collected using IoT (Internet of things) devices. To handle this data using ConvRec model, usus the fine-grained annotation of data points that have the effect of empowering the model to detect relevant traits associated with oxygen stress in fish. It can be therefore ascertained that ConvRec yields high degrees of accuracy of 99.2% and 99.65%, on the “ponds” and “waterx” datasets respectively while the past models can only 98.2% and 98.1% respectively on the same datasets. These results demonstrate that ConvRec is not only promising for estimating the health of fish during oxygen deficiency but also it can take part in reducing the negative impact of low oxygen levels in the water on fish.

Utilizing generative AI for crack detection in the marble industry

Abstract Recently, deep learning algorithms have achieved great accuracies in multiple computer vision tasks by using considerable amounts of annotated data. In the case of marble surface crack detection, the formation of such a dataset is challenging. Currently, there are limited marble crack-related datasets, obstructing the development of robust inspection models for automated industrial quality control. To resolve this problem, we propose a generative Artificial Intelligence (AI) approach, reporting the first synthetic marble crack texture image dataset using four generative adversarial networks (GANs). Each of the final four sub-datasets includes images of different cracks on marble surfaces. The quality of the synthetic datasets is comparatively evaluated; quantitative and qualitative experimental results are provided, highlighting the most efficient generative method. Moreover, to verify the effectiveness and significance of this work, two deep learning segmentation models are used to assess the quality of the synthetic marble crack image datasets. Results indicate that different training strategies using synthetic data could boost the performance of segmentation by up to 6.55% compared to conventional data augmentation techniques. Thus, the first step towards deep learning-based automated marble crack segmentation is delivered, aiming for the simultaneous robotic resin application for healing cracks on-site, in marble-processing industrial settings.

Rapid ecological data collection from 360‐degree imagery using visualisation and immersive sampling in the R pannotator package

Abstract With rapid environmental change occurring globally there is an urgent need for new field survey methods that enable fast and efficient collection of ecological information from impacted systems. Advances in 360‐degree (panospheric) camera technology now allow the collection of geospatially explicit visual records of ecological properties far more cheaply, quickly and efficiently than traditional field survey methods. However, generalisable workflows and software platforms to manage, extract, process and use ecological information from large panospheric image datasets are still lacking. Here we develop a flexible, integrative workflow for rapid acquisition, preparation and extraction of macroscopic data from 360‐degree imagery. We introduce the open‐source R package Panospheric Image Annotator in R (pannotator), which allows the user to visualise and annotate images and extract cropped geocoded sub‐images in a repeatable, systematic way using customisable drop‐down menus and help files. We demonstrate the workflow and pannotator package using panoramic images collected from a study area at Uluṟu‐Kata Tjuṯa National Park in central Australia, which was affected by severe drought and fire in 2018–2020. In this study, 180 images were captured using GoPro Max cameras and imported into the pannotator package for data annotation. We extracted data for three key ecological attributes (plant species distribution, understorey cover and tree health) and show how these data can be used to spatially reconstruct species richness and community structure, plant size class, mortality and burn history. Modern 360‐degree cameras and immersive sampling using the pannotator package now offer a transformative solution for efficiently capturing and extracting macroscopic ecological and biogeographical data from field surveys. More generally, the package may be used to visualise and extract data from any georeferenced panospheric imagery and generate cropped images with embedded geolocation data for use in downstream artificial intelligence/machine learning–based applications.

Identifying protected health information by transformers-based deep learning approach in Chinese medical text.

Purpose: In the context of Chinese clinical texts, this paper aims to propose a deep learning algorithm based on Bidirectional Encoder Representation from Transformers (BERT) to identify privacy information and to verify the feasibility of our method for privacy protection in the Chinese clinical context. Methods: We collected and double-annotated 33,017 discharge summaries from 151 medical institutions on a municipal regional health information platform, developed a BERT-based Bidirectional Long Short-Term Memory Model (BiLSTM) and Conditional Random Field (CRF) model, and tested the performance of privacy identification on the dataset. To explore the performance of different substructures of the neural network, we created five additional baseline models and evaluated the impact of different models on performance. Results: Based on the annotated data, the BERT model pre-trained with the medical corpus showed a significant performance improvement to the BiLSTM-CRF model with a micro-recall of 0.979 and an F1 value of 0.976, which indicates that the model has promising performance in identifying private information in Chinese clinical texts. Conclusions: The BERT-based BiLSTM-CRF model excels in identifying privacy information in Chinese clinical texts, and the application of this model is very effective in protecting patient privacy and facilitating data sharing.

Text classification for evaluating digital technology adoption maturity based on BERT: An evidence of Industrial AI from China

Natural language processing (NLP) models, such as GPT and BERT, are predictable to subvert the research paradigm of technology foresight and innovation management, for their good objectivity, robustness, and efficiency. This paper aims to apply an NLP model based on deep learning to realize the digital technology adoption maturity evaluation. Firstly, a 3-layer evaluation system, with a hierarchy of domain-indicator-level, is proposed. Meanwhile, a dataset on the deployment of Industrial AI in China is collected and provided to our evaluation system. After data annotation by experts with reference to domains and indicators, a BERT model is introduced to execute the multi-label text classification task. The experiment results prove that our high-performance BERT model has the ability to learn from human experts, and then benefits to mitigate biases and reduce cost in evaluation. In the area of Industrial AI deployments, our research points out the digital technologies adoption maturity trends over time, trickle-down effect across regions, and the flying geese pattern between industries.

A survey of open-access datasets for computer vision in precision poultry farming.

Computer vision has progressively advanced precision poultry farming. Despite this substantial increase in research activity, computer vision in precision poultry farming still lacks large-scale, open-access datasets with consistent evaluation metrics and baselines, which makes it challenging to reproduce and validate comparisons of different approaches. Since 2019, several image/video datasets have been published and open-accessed to alleviate the issue of dataset scarcity. However, there is no a dedicated survey summarizing the existing progress. To fill this gap, the objective of this research was to provide the first survey and analysis of the open-access image/video dataset for precision poultry farming. A total of 20 qualified images/video datasets were summarized, including 4 for behavior monitoring, 6 for health status identification, 3 for live performance prediction, 4 for product quality inspection, and 3 for animal trait recognition. Critical points of creating a new image/video dataset, consisting of data acquisition, augmentation, annotation, sharing, and benchmarking, were discussed. The survey provides options for selecting appropriate datasets for model development and optimization while delivering insights into building new datasets for precision poultry farming.

Semi-supervised single-view 3D reconstruction via multi shape prior fusion strategy and self-attention

In the domain of single-view 3D reconstruction, traditional techniques have frequently relied on expensive and time-intensive 3D annotation data. Facing the challenge of annotation acquisition, semi-supervised learning strategies offer an innovative approach to reduce the dependence on labeled data. Despite these developments, the utilization of this learning paradigm in 3D reconstruction tasks remains relatively constrained. In this research, we created an innovative semi-supervised framework for 3D reconstruction that distinctively uniquely introduces a multi shape prior fusion strategy, intending to guide the creation of more realistic object structures. Additionally, to improve the quality of shape generation, we integrated a self-attention module into the traditional decoder. In benchmark tests on the ShapeNet dataset, our method substantially outperformed existing supervised learning methods at diverse labeled ratios of 1%, 10%, and 20%. Moreover, it showcased excellent performance on the real-world Pix3D dataset. Through comprehensive experiments on ShapeNet, our framework demonstrated a 3.3% performance improvement over the baseline. Moreover, stringent ablation studies further confirmed the notable effectiveness of our approach. Our code has been released on https://github.com/NWUzhouwei/SSMP.

Automated segmentation of child-clinician speech in naturalistic clinical contexts.

Computational approaches hold significant promise for enhancing diagnosis and therapy in child and adolescent clinical practice. Clinical procedures heavily depend n vocal exchanges and interpersonal dynamics conveyed through speech. Research highlights the importance of investigating acoustic features and dyadic interactions during child development. However, observational methods are labor-intensive, time-consuming, and suffer from limited objectivity and quantification, hindering translation to everyday care. We propose a novel AI-based system for fully automatic acoustic segmentation of clinical sessions with autistic preschool children. We focused on naturalistic and unconstrained clinical contexts, which are characterized by background noise and data scarcity. Our approach addresses key challenges in the field while remaining non-invasive. We carefully evaluated model performance and flexibility in diverse, challenging conditions by means of domain alignment. Results demonstrated promising outcomes in voice activity detection and speaker diarization. Notably, minimal annotation efforts -just 30 seconds of target data- significantly improved model performance across all tested conditions. Our models exhibit satisfying predictive performance and flexibility for deployment in everyday settings. Automating data annotation in real-world clinical scenarios can enable the widespread exploitation of advanced computational methods for downstream modeling, fostering precision approaches that bridge research and clinical practice.

Universal conditional networks (UniCoN) for multi-age embryonic cartilage segmentation with Sparsely annotated data

Osteochondrodysplasia, affecting 2–3% of newborns globally, is a group of bone and cartilage disorders that often result in head malformations, contributing to childhood morbidity and reduced quality of life. Current research on this disease using mouse models faces challenges since it involves accurately segmenting (precisely delineating) the developing cartilage in 3D micro-CT images of embryonic mice. Tackling this segmentation task with deep learning (DL) methods is laborious due to the big burden of manual image annotation, expensive due to the high acquisition costs of 3D micro-CT images, and difficult due to embryonic cartilage’s complex and rapidly changing shapes. While DL approaches have been proposed to automate cartilage segmentation, most such models have limited accuracy and generalizability, especially across data from different embryonic age groups. To address these limitations, we propose novel DL methods that can be adopted by any DL architectures—including Convolutional Neural Networks (CNNs), Transformers, or hybrid models—which effectively leverage age and spatial information to enhance model performance. Specifically, we propose two new mechanisms, one conditioned on discrete age categories and the other on continuous image crop locations, to enable an accurate representation of cartilage shape changes across ages and local shape details throughout the cranial region. Extensive experiments on multi-age cartilage segmentation datasets show significant and consistent performance improvements when integrating our conditional modules into popular DL segmentation architectures. On average, we achieve a 1.7% Dice score increase with minimal computational overhead and a 7.5% improvement on unseen data. These results highlight the potential of our approach for developing robust, universal models capable of handling diverse datasets with limited annotated data, a key challenge in DL-based medical image analysis.

Improving the Reliability of Evolutionary Algorithm for Complex Detection in Noisy Protein-Protein Interaction Networks

Evolutionary algorithms are better than heuristic algorithms at finding protein complexes in protein-protein interaction networks (PPINs). Many of these algorithms depend on their standard frameworks, which are based on topology. Further, many of these algorithms have been exclusively examined on networks with only reliable interaction data. The main objective of this paper is to extend the design of the canonical and topological-based evolutionary algorithms suggested in the literature to cope with noisy PPINs. The design of the evolutionary algorithm is extended based on the functional domain of the proteins rather than on the topological domain of the PPIN. The gene ontology annotation in each molecular function, biological process, and cellular component is used to get the functional domain. The reliability of the proposed algorithm is examined against the algorithms proposed in the literature. To this end, a yeast protein-protein interaction dataset is used in the assessment of the final quality of the algorithms. To make fake negative controls of PPIs that are wrongly informed and are linked to the high-throughput interaction data, different noisy PPINs are created. The noisy PPINs are synthesized with a different and increasing percentage of misinformed PPIs. The results confirm the effectiveness of the extended evolutionary algorithm design to utilize the biological knowledge of the gene ontology. Feeding EA design with GO annotation data improves reliability and produces more accurate detection results than the counterpart algorithms.

Ethical Concerns in Large-Scale Data Labeling and Usage: Challenges and Solutions

Abstract: The rapid adoption of artificial intelligence (AI) and machine learning (ML) has created an unprecedented demand for high-quality labeled data. Large-scale data labeling, a critical component of AI system development, often involves vast datasets sourced from diverse populations and annotated using a combination of automated processes and human labor. However, the ethical challenges associated with these practices have gained significant attention. This paper explores key ethical concerns in large-scale data labeling and usage, focusing on four critical areas: bias, privacy, labor practices, and transparency. Bias in labeled data, arising from the inherent subjectivity of annotators and the unrepresentative nature of many datasets, exacerbates the risk of unfair or discriminatory outcomes in AI applications. Privacy violations occur when sensitive information is collected or used without proper consent, often challenging the effectiveness of anonymization techniques. Furthermore, the reliance on crowdsourced labor for data annotation raises concerns about worker exploitation, low compensation, and the mental toll of labeling sensitive or explicit content. Lastly, the lack of transparency and accountability in data collection and labeling processes undermines public trust and ethical standards. Through a comprehensive review of existing practices, this paper highlights real-world case studies and controversies, including biased datasets and privacy violations. Current technological and policy-driven solutions—such as privacy-preserving techniques, labor reforms, and bias mitigation algorithms—are critically examined. Finally, the paper discusses the challenges of implementing these solutions at scale and identifies future research directions. By addressing these concerns, this work aims to promote more equitable, transparent, and ethical practices in the lifecycle of AI data management.

Frames and Triggers of Extreme Speech: The Case of Transphobia

Scholars have begun to study online representations of transgender people (Craig et al. 2015; McInroy and Craig 2015) and to research discriminatory language aimed specifically at trans people, or ‘transphobic language’ (Baker 2004; Turner et al. 2009; Colliver et al. 2019; Zottola 2021). And yet, despite the fact that homophobic discourse/ anti-LGBTQ+ discourse is an established field of research, transphobic online comments have not yet been addressed from a critical discourse analysis perspective, except for the Colliver et al. 2019 study, which adopted a critical discursive psychology framework. In the present study, therefore, we examine the discursive strategies used to construct the othering of these specific non-heteronormative identities and the counter speech used to deconstruct them. More precisely, we aim to understand what triggers transphobic comments, what argumentative frames are most used, what counter-speech strategies are present in the online debates, and how effective they are. We answer these research questions based on online data focused on transphobic speech collected in 2017. Our methodology includes analyzing annotated computational data, followed by discourse analysis and consideration of theoretical results from the ARENAS project (2023-2027). To begin we will define the fundamental concepts used in this study: what distinguishes hate speech from extreme speech, and the frames and triggers that we have identified in the context of the present research. After a brief literature review, we present our data and the quantitative results, after which we analyze the main frames and triggers of transphobic speech. To conclude, we examine the counter speech used to deconstruct the discriminatory argumentative frames.

GAN-CITE: leveraging semi-supervised generative adversarial networks for citation function classification with limited data

Abstract Citation function analysis is crucial to understanding how cited literature contributes to the overall discourse and meaning conveyed in scientific publications. Citation functions serve diverse roles that must be accurately identified and categorized. Still, the field of citation function analysis faces challenges due to limited labeled data and the complexity of defining and categorizing citation functions, which require expertise and a deep understanding of scientific literature. This limitation results in imprecise identification and categorization of citation functions, emphasizing the need for further advancements to improve the accuracy and reliability of citation function analysis. This paper proposes GAN-CITE, a novel framework employing semi-supervised learning techniques to address these limitations. Its primary objective is to efficiently leverage available unlabeled data by combining generative adversarial networks (GANs) and the language model to incorporate substantial data representations from unlabeled data sources. Our study demonstrates that GAN-CITE outperforms both supervised and semi-supervised state-of-the-art models in limited data settings, namely 10%, 20%, and 30% of the total labeled data. We also examine its performance in insufficient and imbalanced labeled data situations, as well as the potential of unlabeled data utilization. These findings highlight the success of generative adversarial networks in enhancing citation function classification and their applications in digital libraries that require precise citation function categorization, such as trend analysis and impact quantification, under limited annotated data.

“Everybody knows what a pothole is”: representations of work and intelligence in AI practice and governance

Abstract In this paper, we empirically and conceptually examine how distributed human–machine networks of labour comprise a form of underlying intelligence within Artificial Intelligence (AI), considering the implications of this for Responsible Artificial Intelligence (R-AI) innovation. R-AI aims to guide AI research, development and deployment in line with certain normative principles, for example fairness, privacy, and explainability; notions implicitly shaped by comparisons of AI with individualised notions of human intelligence. However, as critical scholarship on AI demonstrates, this is a limited framing of the nature of intelligence, both of humans and AI. Furthermore, it dismisses the skills and labour central to developing AI systems, involving a distributed network of human-directed practices and reasoning. We argue that inequities in the agency and recognition of different types of practitioners across these networks of AI development have implications beyond RAI, with narrow framings concealing considerations which are important within broader discussions of AI intelligence. Drawing from interactive workshops conducted with AI practitioners, we explore practices of data acquisition, cleaning, and annotation, as the point where practitioners interface with domain experts and data annotators. Despite forming a crucial part of AI design and development, this type of data work is frequently framed as a tedious, unskilled, and low-value process. In exploring these practices, we examine the political role of the epistemic framings that underpin AI development and how these framings can shape understandings of distributed intelligence, labour practices, and annotators’ agency within data structures. Finally, we reflect on the implications of our findings for developing more participatory and equitable approaches to machine learning applications in the service of R-AI.

Multimodal Augmented Reality System for Real-Time Roof Type Recognition and Visualization on Mobile Devices

The utilization of augmented reality (AR) is becoming increasingly prevalent in the integration of virtual reality (VR) elements into the tangible reality of the physical world. It facilitates a more straightforward comprehension of the interconnections, interdependencies, and spatial context of data. Furthermore, the presentation of analyses and the combination of spatial data with annotated data are facilitated. This is particularly evident in the context of mobile applications, where the combination of real-world and virtual imagery facilitates enhances visualization. This paper presents a proposal for the development of a multimodal system that is capable of identifying roof types in real time and visualizing them in AR on mobile devices. The current approach to roof identification is based on data made available by public administrations in an open-source format, including orthophotos and building contours. Existing computer processing technologies have been employed to generate objects representing the shapes of building masses, and in particular, the shape of roofs, in three-dimensional (3D) space. The system integrates real-time data obtained from multiple sources and is based on a mobile application that enables the precise positioning and detection of the recipient’s viewing direction (pose estimation) in real time. The data were integrated and processed in a Docker container system, which ensured the scalability and security of the solution. The multimodality of the system is designed to enhance the user’s perception of the space and facilitate a more nuanced interpretation of its intricacies. In its present iteration, the system facilitates the extraction and classification/generalization of two categories of roof types (gable and other) from aerial imagery through the utilization of deep learning methodologies. The outcomes achieved suggest considerable promise for the advancement and deployment of the system in domains pertaining to architecture, urban planning, and civil engineering.

The role of SASP in ischemic stroke: a deep dive into cellular mechanisms

BackgroundThe escalating incidence of ischemic stroke (IS) exerts a heavy toll on global health. Aging, a prominent risk factor, implicates the senescence-associated secretory phenotype (SASP) in IS pathogenesis. We postulated that alterations in SASP-related factor expression during IS correlate with remodeling of intercellular interaction networks and disease advancement. The present study endeavored to preliminarily dissect the SASP-IS nexus via combined bulk and single-cell transcriptome analysis.MethodsAggregated expression profiles from human peripheral blood bulk chips and MCAO mouse single-cell sequencing data, followed by SASP gene analysis. Executed protein interaction network and enrichment assays. Investigated immune infiltration in stroke patients, managed quality control and annotation of single-cell data, cherry-picked central cells based on SASP scores, unearthed essential genes via enrichment analysis, conducted pseudo-time and intercellular communication studies, and prognosticated drugs for hub genes. Finally, authenticated core gene expression in serum of MCAO and Sham rats using real-time fluorescent polymerase chain reaction (RT-qPCR).ResultsFourteen hub genes were discerned. Seven cell types were annotated in MCAO mouse peripheral blood single-cell data. Basophils exhibited the highest SASP scores, with Lcp1 upregulated and Ccl3 downregulated in basophils of the MCAO group. Enrichment analysis divulged a significant association of Ccl3 with the cell apoptosis pathway and Lcp1 with immune responses. The Ccl3 gene is pivotal in basophils and basophil-neutrophil crosstalk. Additionally, we forecasted nagrestipen’s regulatory function on Ccl3. RT-qPCR demonstrated a marked elevation in Lcp1 mRNA and a pronounced reduction in Ccl3 in the MCAO group relative to the Sham group.ConclusionThe Ccl3 gene in basophils and its immune cell interaction is a linchpin in the IS immune microenvironment. Ccl3 and Lcp1 might potentially modulate IS progression by influencing SASP, proffering novel prospects for IS clinical diagnosis and treatment.

Flood change detection model based on an improved U-net network and multi-head attention mechanism

This work aims to improve the accuracy and efficiency of flood disaster monitoring, including monitoring before, during, and after the flood, to achieve accurate extraction of flood disaster change information. A modified U-Net network model, incorporating the Transformer multi-head attention mechanism (TM), is developed specifically for the characteristics of Synthetic Aperture Radar (SAR) images. By integrating the TM, the model effectively prioritizes image regions relevant to flood disasters. The model is trained on a substantial volume of annotated SAR image data, and its performance is assessed using metrics such as loss function, accuracy, and precision. Experimental findings demonstrate significant improvements in loss value, accuracy, and precision compared to existing models. Specifically, the accuracy of the model algorithm in this work reaches 95.52%, marking a 3.46% improvement over the baseline U-Net network. Additionally, the developed model achieves an accuracy of 90.11% while maintaining a loss value of approximately 0.59, whereas other model algorithms exceed a loss value of 0.74. Thus, this work not only introduces a novel technical approach for flood disaster monitoring but also has the potential to enhance disaster response procedures and provide scientific evidence for disaster management and risk assessment processes.

Heterogeneous Graph Neural Networks using Self-supervised Reciprocally Contrastive Learning

Heterogeneous graph neural network (HGNN) is a popular technique for modeling and analyzing heterogeneous graphs. Most existing HGNN-based approaches are supervised or semi-supervised learning methods requiring graphs to be annotated, which is costly and time-consuming. Self-supervised contrastive learning has been proposed to address the problem of requiring annotated data by mining intrinsic properties in the given data. However, the existing contrastive learning methods are not suitable for heterogeneous graphs because they construct contrastive views only based on data perturbation or pre-defined structural properties (e.g., meta-path) in graph data while ignoring noises in node attributes and graph topologies. We develop a robust heterogeneous graph contrastive learning approach, namely HGCL, which introduces two views on respective guidances of node attributes and graph topologies and integrates and enhances them by a reciprocally contrastive mechanism to better model heterogeneous graphs. In this new approach, we adopt distinct but suitable attribute and topology fusion mechanisms in the two views, which are conducive to mining relevant information in attributes and topologies separately. We further use both attribute similarity and topological correlation to construct high-quality contrastive samples. Extensive experiments on four large real-world heterogeneous graphs demonstrate the superiority and robustness of HGCL over several state-of-the-art methods.