Fractal feature selection model for enhancing high-dimensional biological problems.

The integration of biology, computer science, and statistics has given rise to the interdisciplinary field of bioinformatics, which aims to decode biological intricacies. It produces extensive and diverse features, presenting an enormous challenge in classifying bioinformatic problems. Therefore, an intelligent bioinformatics classification system must select the most relevant features to enhance machine learning performance. This paper proposes a feature selection model based on the fractal concept to improve the performance of intelligent systems in classifying high-dimensional biological problems. The proposed fractal feature selection (FFS) model divides features into blocks, measures the similarity between blocks using root mean square error (RMSE), and determines the importance of features based on low RMSE. The proposed FFS is tested and evaluated over ten high-dimensional bioinformatics datasets. The experiment results showed that the model significantly improved machine learning accuracy. The average accuracy rate was 79% with full features in machine learning algorithms, while FFS delivered promising results with an accuracy rate of 94%.

Using protein language models for protein interaction hot spot prediction with limited data.

BACKGROUND: Protein language models, inspired by the success of large language models in deciphering human language, have emerged as powerful tools for unraveling the intricate code of life inscribed within protein sequences. They have gained significant attention for their promising applications across various areas, including the sequence-based prediction of secondary and tertiary protein structure, the discovery of new functional protein sequences/folds, and the assessment of mutational impact on protein fitness. However, their utility in learning to predict protein residue properties based on scant datasets, such as protein-protein interaction (PPI)-hotspots whose mutations significantly impair PPIs, remained unclear. Here, we explore the feasibility of using protein language-learned representations as features for machine learning to predict PPI-hotspots using a dataset containing 414 experimentally confirmed PPI-hotspots and 504 PPI-nonhot spots. RESULTS: Our findings showcase the capacity of unsupervised learning with protein language models in capturing critical functional attributes of protein residues derived from the evolutionary information encoded within amino acid sequences. We show that methods relying on protein language models can compete with methods employing sequence and structure-based features to predict PPI-hotspots from the free protein structure. We observed an optimal number of features for model precision, suggesting a balance between information and overfitting. CONCLUSIONS: This study underscores the potential of transformer-based protein language models to extract critical knowledge from sparse datasets, exemplified here by the challenging realm of predicting PPI-hotspots. These models offer a cost-effective and time-efficient alternative to traditional experimental methods for predicting certain residue properties. However, the challenge of explaining why specific features are important for determining certain residue properties remains.

Genetic variations and gene expression profiles of Rice Black-streaked dwarf virus (RBSDV) in different host plants and insect vectors: insights from RNA-Seq analysis.

Rice black-streaked dwarf virus (RBSDV) is an etiological agent of a destructive disease infecting some economically important crops from the Gramineae family in Asia. While RBSDV causes high yield losses, genetic characteristics of replicative viral populations have not been investigated within different host plants and insect vectors. Herein, eleven publicly available RNA-Seq datasets from Chinese RBSDV-infected rice, maize, and viruliferous planthopper (Laodelphax striatellus) were obtained from the NCBI database. The patterns of SNP and RNA expression profiles of expected RBSDV populations were analyzed by CLC Workbench 20 and Geneious Prime software. These analyses discovered 2,646 mutations with codon changes in RBSDV whole transcriptome and forty-seven co-mutated hotspots with high variant frequency within the crucial regions of S5-1, S5-2, S6, S7-1, S7-2, S9, and S10 open reading frames (ORFs) which are responsible for some virulence and host range functions. Moreover, three joint mutations are located on the three-dimensional protein of P9-1. The infected RBSDV-susceptible rice cultivar KTWYJ3 and indigenous planthopper datasets showed more co-mutated hotspot numbers than others. Our analyses showed the expression patterns of viral genomic fragments varied depending on the host type. Unlike planthopper, S5-1, S2, S6, and S9-1 ORFs, respectively had the greatest read numbers in host plants; and S5-2, S9-2, and S7-2 were expressed in the lowest level. These findings underscore virus/host complexes are effective in the genetic variations and gene expression profiles of plant viruses. Our analysis revealed no evidence of recombination events. Interestingly, the negative selection was observed at 12 RBSDV ORFs, except for position 1015 in the P1 protein, where a positive selection was detected. The research highlights the potential of SRA datasets for analysis of the virus cycle and enhances our understanding of RBSDV's genetic diversity and host specificity.

New approaches to predict the effect of co-occurring variants on protein characteristics.

Predicting the effect of a mutated gene before the onset of symptoms of genetic diseases would greatly facilitate diagnosis and potentiate early intervention. There have been myriad attempts to predict the effects of single-nucleotide variants. However, the applicability of these efforts does not scale to co-occurring variants. Furthermore, an increasing number of protein therapeutics contain co-occurring nucleotide variations, adding uncertainty during development to the safety and efficiency of these drugs. Co-occurring nucleotide variants may often have synergistic, additive, or antagonistic effects on protein attributes, further complicating the task of outcome prediction. We tested four models based on the cooperative and antagonistic effects of co-occurring variants to predict pathogenicity and effectiveness of protein therapeutics. A total of 30 attributes, including amino acid and nucleotide features, as well as existing single-variant effect prediction tools, were considered on the basis of previous studies on single-nucleotide variants. Importantly, the effects of synonymous variants, often seen in protein therapeutics, were also included in our models. We used 12 datasets of people with monogenic diseases and controls with co-occurring genetic variants to evaluate the accuracy of our models, accomplishing a degree of accuracy comparable to that of prediction tools for single-nucleotide variants. More importantly, our framework is generalizable to new, well-curated datasets of monogenic diseases and new variant scoring tools. This approach successfully assists in addressing the challenging task of predicting the effect of co-occurring variants on pathogenicity and protein effectiveness and is applicable for a wide range of protein therapeutics and genetic diseases.

Dataset factors associated with age-related changes in brain structure and function in neurodevelopmental conditions.

With brain structure and function undergoing complex changes throughout childhood and adolescence, age is a critical consideration in neuroimaging studies, particularly for those of individuals with neurodevelopmental conditions. However, despite the increasing use of large, consortium-based datasets to examine brain structure and function in neurotypical and neurodivergent populations, it is unclear whether age-related changes are consistent between datasets and whether inconsistencies related to differences in sample characteristics, such as demographics and phenotypic features, exist. To address this, we built models of age-related changes of brain structure (regional cortical thickness and regional surface area; N = 1218) and function (resting-state functional connectivity strength; N = 1254) in two neurodiverse datasets: the Province of Ontario Neurodevelopmental Network and the Healthy Brain Network. We examined whether deviations from these models differed between the datasets, and explored whether these deviations were associated with demographic and clinical variables. We found significant differences between the two datasets for measures of cortical surface area and functional connectivity strength throughout the brain. For regional measures of cortical surface area, the patterns of differences were associated with race/ethnicity, while for functional connectivity strength, positive associations were observed with head motion. Our findings highlight that patterns of age-related changes in the brain may be influenced by demographic and phenotypic characteristics, and thus future studies should consider these when examining or controlling for age effects in analyses.

GLOBE: a contrastive learning-based framework for integrating single-cell transcriptome datasets.

Integration of single-cell transcriptome datasets from multiple sources plays an important role in investigating complex biological systems. The key to integration of transcriptome datasets is batch effect removal. Recent methods attempt to apply a contrastive learning strategy to correct batch effects. Despite their encouraging performance, the optimal contrastive learning framework for batch effect removal is still under exploration. We develop an improved contrastive learning-based batch correction framework, GLOBE. GLOBE defines adaptive translation transformations for each cell to guarantee the stability of approximating batch effects. To enhance the consistency of representations alignment, GLOBE utilizes a loss function that is both hardness-aware and consistency-aware to learn batch effect-invariant representations. Moreover, GLOBE computes batch-corrected gene matrix in a transparent approach to support diverse downstream analysis. Benchmarking results on a wide spectrum of datasets show that GLOBE outperforms other state-of-the-art methods in terms of robust batch mixing and superior conservation of biological signals. We further apply GLOBE to integrate two developing mouse neocortex datasets and show GLOBE succeeds in removing batch effects while preserving the contiguous structure of cells in raw data. Finally, a comprehensive study is conducted to validate the effectiveness of GLOBE.

JEBIN: analyzing gene co-expressions across multiple datasets by joint network embedding.

The inference of gene co-expression associations is one of the fundamental tasks for large-scale transcriptomic data analysis. Due to the high dimensionality and high noises in transcriptomic data, it is difficult to infer stable gene co-expression associations from single dataset. Meta-analysis of multisource data can effectively tackle this problem. We proposed Joint Embedding of multiple BIpartite Networks (JEBIN) to learn the low-dimensional consensus representation for genes by integrating multiple expression datasets. JEBIN infers gene co-expression associations in a nonlinear and global similarity manner and can integrate datasets with different distributions in linear time complexity with the gene and total sample size. The effectiveness and scalability of JEBIN were verified by simulation experiments, and its superiority over the commonly used integration methods was proved by three indexes on real biological datasets. Then, JEBIN was applied to study the gene co-expression patterns of hepatocellular carcinoma (HCC) based on multiple expression datasets of HCC and adjacent normal tissues, and further on latest HCC single-cell RNA-seq data. Results show that gene co-expressions are highly different between bulk and single-cell datasets. Finally, many differentially co-expressed ligand-receptor pairs were discovered by comparing HCC with adjacent normal data, providing candidate HCC targets for abnormal cell-cell communications.

Perspective: use and reuse of NMR-based metabolomics data: what works and what remains challenging.

BACKGROUND: The National Cancer Institute issued a Request for Information (RFI; NOT-CA-23-007) in October 2022, soliciting input on using and reusing metabolomics data. This RFI aimed to gather input on best practices for metabolomics data storage, management, and use/reuse. AIM OF REVIEW: The nuclear magnetic resonance (NMR) Interest Group within the Metabolomics Association of North America (MANA) prepared a set of recommendations regarding the deposition, archiving, use, and reuse of NMR-based and, to a lesser extent, mass spectrometry (MS)-based metabolomics datasets. These recommendations were built on the collective experiences of metabolomics researchers within MANA who are generating, handling, and analyzing diverse metabolomics datasets spanning experimental (sample handling and preparation, NMR/MS metabolomics data acquisition, processing, and spectral analyses) to computational (automation of spectral processing, univariate and multivariate statistical analysis, metabolite prediction and identification, multi-omics data integration, etc.) studies. KEY SCIENTIFIC CONCEPTS OF REVIEW: We provide a synopsis of our collective view regarding the use and reuse of metabolomics data and articulate several recommendations regarding best practices, which are aimed at encouraging researchers to strengthen efforts toward maximizing the utility of metabolomics data, multi-omics data integration, and enhancing the overall scientific impact of metabolomics studies.

Too hot to handle? An urgent need to understand climate change impacts on the biogeochemistry of tropical coastal waters.

Tropical regions contain ecologically and socio-economically important habitats, and are home to about 3.8 billion people, many of which directly depend on tropical coastal waters for their well-being. At the basis of these ecosystems are biogeochemical processes. Climate change is expected to have a greater impact in the tropics compared to temperate regions because of the relatively stable environmental conditions found there. However, it was surprising to find only 660 research articles published focusing on the impact of climate change on the biogeochemistry of coastal tropical waters compared to 4823 for temperate waters. In this perspective, we highlight important topics in need of further research. Specifically, we suggest that in tropical regions compared to temperate counterparts climate change stressors will be experienced differently, that organisms have a lower acclimation capacity, and that long-term baseline biogeochemical datasets useful for quantifying future changes are lacking. The low number of research papers on the impacts of climate change in coastal tropical regions is likely due to a mix of reasons including limited resources for research and limited number of long time series in many developing tropical countries. Finally, we propose some action points that we hope will stimulate more studies in tropical coastal waters.

Gray matter atrophy is constrained by normal structural brain network architecture in depression.

BACKGROUND: There is growing evidence that gray matter atrophy is constrained by normal brain network (or connectome) architecture in neuropsychiatric disorders. However, whether this finding holds true in individuals with depression remains unknown. In this study, we aimed to investigate the association between gray matter atrophy and normal connectome architecture at individual level in depression. METHODS: In this study, 297 patients with depression and 256 healthy controls (HCs) from two independent Chinese dataset were included: a discovery dataset (105 never-treated first-episode patients and matched 130 HCs) and a replication dataset (106 patients and matched 126 HCs). For each patient, individualized regional atrophy was assessed using normative model and brain regions whose structural connectome profiles in HCs most resembled the atrophy patterns were identified as putative epicenters using a backfoward stepwise regression analysis. RESULTS: In general, the structural connectome architecture of the identified disease epicenters significantly explained 44% (±16%) variance of gray matter atrophy. While patients with depression demonstrated tremendous interindividual variations in the number and distribution of disease epicenters, several disease epicenters with higher participation coefficient than randomly selected regions, including the hippocampus, thalamus, and medial frontal gyrus were significantly shared by depression. Other brain regions with strong structural connections to the disease epicenters exhibited greater vulnerability. In addition, the association between connectome and gray matter atrophy uncovered two distinct subgroups with different ages of onset. CONCLUSIONS: These results suggest that gray matter atrophy is constrained by structural brain connectome and elucidate the possible pathological progression in depression.

A Survey of Publicly Available MRI Datasets for Potential Use in Artificial Intelligence Research.

Artificial intelligence (AI) has the potential to bring transformative improvements to the field of radiology; yet, there are barriers to widespread clinical adoption. One of the most important barriers has been access to large, well-annotated, widely representative medical image datasets, which can be used to accurately train AI programs. Creating such datasets requires time and expertise and runs into constraints around data security and interoperability, patient privacy, and appropriate data use. Recognizing these challenges, several institutions have started curating and providing publicly available, high-quality datasets that can be accessed by researchers to advance AI models. The purpose of this work was to review the publicly available MRI datasets that can be used for AI research in radiology. Despite being an emerging field, a simple internet search for open MRI datasets presents an overwhelming number of results. Therefore, we decided to create a survey of the major publicly accessible MRI datasets in different subfields of radiology (brain, body, and musculoskeletal), and list the most important features of value to the AI researcher. To complete this review, we searched for publicly available MRI datasets and assessed them based on several parameters (number of subjects, demographics, area of interest, technical features, and annotations). We reviewed 110 datasets across sub-fields with 1,686,245 subjects in 12 different areas of interest ranging from spine to cardiac. This review is meant to serve as a reference for researchers to help spur advancements in the field of AI for radiology. LEVEL OF EVIDENCE: Level 4 TECHNICAL EFFICACY: Stage 6.

Implications of Bias in Artificial Intelligence: Considerations for Cardiovascular Imaging.

PURPOSE OF REVIEW: Bias in artificial intelligence (AI) models can result in unintended consequences. In cardiovascular imaging, biased AI models used in clinical practice can negatively affect patient outcomes. Biased AI models result from decisions made when training and evaluating a model. This paper is a comprehensive guide for AI development teams to understand assumptions in datasets and chosen metrics for outcome/ground truth, and how this translates to real-world performance for cardiovascular disease (CVD). RECENT FINDINGS: CVDs are the number one cause of mortality worldwide; however, the prevalence, burden, and outcomes of CVD vary across gender and race. Several biomarkers are also shown to vary among different populations and ethnic/racial groups. Inequalities in clinical trial inclusion, clinical presentation, diagnosis, and treatment are preserved in health data that is ultimately used to train AI algorithms, leading to potential biases in model performance. Despite the notion that AI models themselves are biased, AI can also help to mitigate bias (e.g., bias auditing tools). In this review paper, we describe in detail implicit and explicit biases in the care of cardiovascular disease that may be present in existing datasets but are not obvious to model developers. We review disparities in CVD outcomes across different genders and race groups, differences in treatment of historically marginalized groups, and disparities in clinical trials for various cardiovascular diseases and outcomes. Thereafter, we summarize some CVD AI literature that shows bias in CVD AI as well as approaches that AI is being used to mitigate CVD bias.

The PACIFIC ontology for heterogeneous data management in cardiology.

With the emergence of health data warehouses and major initiatives to collect and analyze multi-modal and multisource data, data organization becomes central. In the PACIFIC-PRESERVED (PhenomApping, ClassIFication, and Innovation for Cardiac Dysfunction - Heart Failure with PRESERVED LVEF Study, NCT04189029) study, a data driven research project aiming at redefining and profiling the Heart Failure with preserved Ejection Fraction (HFpEF), an ontology was developed by different data experts in cardiology to enable better data management in a complex study context (multisource, multiformat, multimodality, multipartners). The PACIFIC ontology provides a cardiac data management framework for the phenomapping of patients. It was built upon the BMS-LM (Biomedical Study -Lifecycle Management) core ontology and framework, proposed in a previous work to ensure data organization and provenance throughout the study lifecycle (specification, acquisition, analysis, publication). The BMS-LM design pattern was applied to the PACIFIC multisource variables. In addition, data was structured using a subset of MeSH headings for diseases, technical procedures, or biological processes, and using the Uberon ontology anatomical entities. A total of 1372 variables were organized and enriched with annotations and description from existing ontologies and taxonomies such as LOINC to enable later semantic interoperability. Both, data structuring using the BMS-LM framework, and its mapping with published standards, foster interoperability of multimodal cardiac phenomapping datasets.

Burden and distribution of mortality due to sepsis and severe infection in children and adolescents in Aotearoa/New Zealand.

AIM: The aims of this research were to determine the mortality from sepsis and severe infection in the paediatric and adolescent populations of Aotearoa/New Zealand, and to determine the distribution of mortality by sub-populations. METHODS: We used three different methods to identify deaths from sepsis and severe infection and compared the groups: All deaths primarily coded with any ICD-10-AM code relating to sepsis; The presence of A40, A41 and P36 in any cause of death field; Deaths due to pneumonia and meningitis. Cases were selected from a national mortality database, with cause of death as ascribed in the national mortality collection for the years 2002-2020 inclusive. Overall sepsis and severe infection rates were calculated from the sum of unique cases from all three methods for determining sepsis and severe infection cases. RESULTS: Substantially different results were obtained depending on the method of identifying cases. In total, 577 deaths due to sepsis and severe infection were detected, with an overall rate of 1.99/100 000 age-specific population and statistically significant disparity by ethnic grouping. Rates were highest in post-neonatal infants at 22.7 per 100 000, regardless of the method of identification. CONCLUSIONS: There is a considerable opportunity to improve the mortality from sepsis and severe infection in children and young people. The ethnic disparities described in this paper show the need to ensure a high level of care for those most marginalised in society through the development and provision of systems and structures that meet, rather than fail to meet need.

A patient safety knowledge graph supporting vaccine product development.

BACKGROUND: Knowledge graphs are well-suited for modeling complex, unstructured, and multi-source data and facilitating their analysis. During the COVID-19 pandemic, adverse event data were integrated into a knowledge graph to support vaccine safety surveillance and nimbly respond to urgent health authority questions. Here, we provide details of this post-marketing safety system using public data sources. In addition to challenges with varied data representations, adverse event reporting on the COVID-19 vaccines generated an unprecedented volume of data; an order of magnitude larger than adverse events for all previous vaccines. The Patient Safety Knowledge Graph (PSKG) is a robust data store to accommodate the volume of adverse event data and harmonize primary surveillance data sources. METHODS: We designed a semantic model to represent key safety concepts. We built an extract-transform-load (ETL) data pipeline to parse and import primary public data sources; align key elements such as vaccine names; integrated the Medical Dictionary for Regulatory Activities (MedDRA); and applied quality metrics. PSKG is deployed in a Neo4J graph database, and made available via a web interface and Application Programming Interfaces (APIs). RESULTS: We import and align adverse event data and vaccine exposure data from 250 countries on a weekly basis, producing a graph with 4,340,980 nodes and 30,544,475 edges as of July 1, 2022. PSKG is used for ad-hoc analyses and periodic reporting for several widely available COVID-19 vaccines. Analysis code using the knowledge graph is 80% shorter than an equivalent implementation written entirely in Python, and runs over 200 times faster. CONCLUSIONS: Organizing safety data into a concise model of nodes, properties, and edge relationships has greatly simplified analysis code by removing complex parsing and transformation algorithms from individual analyses and instead managing these centrally. The adoption of the knowledge graph transformed how the team answers key scientific and medical questions. Whereas previously an analysis would involve aggregating and transforming primary datasets from scratch to answer a specific question, the team can now iterate easily and respond as quickly as requests evolve (e.g., "Produce vaccine-X safety profile for adverse event-Y by country instead of age-range").

Study on differentially expressed genes and participating pathways of ectopic endometrium in adenomyosis patients with different data sets.

BACKGROUND: Adenomyosis is a benign uterine disease and affected patients present with symptoms such as menorrhagia, chronic pelvic pain, abnormal uterine bleeding, and infertility. However, the specific mechanisms by which adenomyosis occurs need to be further studied. OBJECTIVE: Dataset of adenomyosis from our hospital and a public database were analyzed using bioinformatics. Corresponding differentially expressed genes (DEGs) and gene enrichment were detected to explore potential genetic adenomyosis targets. METHODS: Clinical data on adenomyosis were accessed based on the pathological specimens of patients with adenomyosis obtained from the Shengjing Hospital. R software was used to screen for DEGs, and volcano and cluster maps were drawn. Adenomyosis datasets (GSE74373) were downloaded from the GEO database. GEO2R online tool was used to screen for DEGs between adenomyosis and normal controls. Genes with P < 0.01 and |logFC| >1 were selected as DEGs. DAVID software was used for functional and pathway enrichment analyses. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed on common DEGs to obtain descriptions of the genes. The online database STRING was used for interaction gene retrieval. Moreover, Cytoscape software was used to construct a protein-protein interaction (PPI) network map for common DEGs to visualize potential gene interactions and screen the hub genes. RESULTS: A total of 845 DEGs were identified in the dataset obtained from Shengjing Hospital. A total of 175 genes were downregulated, and 670 genes were upregulated. In the GSE74373 database, 1679 genes were differentially expressed, 916 genes were downregulated, and 763 genes were upregulated. A total of 40 downregulated and 148 upregulated common DEGs showed potential gene interactions. The top ten upregulated hub genes were CDH1, EPCAM, CLDN7, ESRP1, RAB25, SPINT1, PKP3, TJP3, GRHL2, and CDKN2A. CONCLUSION: Genes involved in tight junction may be key in the development of adenomyosis and may provide a potential treatment strategy for adenomyosis.

Investigation of Security Threat Datasets for Intra- and Inter-Vehicular Environments.

Vehicular networks have become a critical component of modern transportation systems by facilitating communication between vehicles and infrastructure. Nonetheless, the security of such networks remains a significant concern, given the potential risks associated with cyberattacks. For this purpose, artificial intelligence approaches have been explored to enhance the security of vehicular networks. Using artificial intelligence algorithms to analyze large datasets can enable the early identification and mitigation of potential threats. However, developing and testing effective artificial-intelligence-based solutions for vehicular networks necessitates access to diverse datasets that accurately capture the various security challenges and attack scenarios in this context. In light of this, the present survey comprehensively examines the vehicular network environment, the associated security issues, and existing datasets. Specifically, we begin with a general overview of the vehicular network environment and its security challenges. Following this, we introduce an innovative taxonomy designed to classify datasets pertinent to vehicular network security and analyze key features of these datasets. The survey concludes with a tailored guide aimed at researchers in the vehicular network domain. This guide offers strategic advice on selecting the most appropriate datasets for specific research scenarios in the field.

PSA-FL-CDM: A Novel Federated Learning-Based Consensus Model for Post-Stroke Assessment.

The rapid development of Internet of Things (IoT) technologies and the potential benefits of employing the vast datasets generated by IoT devices, including wearable sensors and camera systems, has ushered in a new era of opportunities for enhancing smart rehabilitation in various healthcare systems. Maintaining patient privacy is paramount in healthcare while providing smart insights and recommendations. This study proposed the adoption of federated learning to develop a scalable AI model for post-stroke assessment while protecting patients' privacy. This research compares the centralized (PSA-MNMF) model performance with the proposed scalable federated PSA-FL-CDM model for sensor- and camera-based datasets. The computational time indicates that the federated PSA-FL-CDM model significantly reduces the execution time and attains comparable performance while preserving the patient's privacy. Impact Statement-This research introduces groundbreaking contributions to stroke assessment by successfully implementing federated learning for the first time in this domain and applying consensus models in each node. It enables collaborative model training among multiple nodes or clients while ensuring the privacy of raw data. The study explores eight different clustering methods independently on each node, revolutionizing data organization based on similarities in stroke assessment. Additionally, the research applies the centralized PSA-MNMF consensus clustering technique to each client, resulting in more accurate and robust clustering solutions. By utilizing the FedAvg federated learning algorithm strategy, locally trained models are combined to create a global model that captures the collective knowledge of all participants. Comparative performance measurements and computational time analyses are conducted, facilitating a fair evaluation between centralized and federated learning models in stroke assessment. Moreover, the research extends beyond a single type of database by conducting experiments on two distinct datasets, wearable and camera-based, broadening the understanding of the proposed methods across different data modalities. These contributions develop stroke assessment methodologies, enabling efficient collaboration and accurate consensus clustering models and maintaining data privacy.

Infrared Dim Small Target Detection Networks: A Review.

In recent years, with the rapid development of deep learning and its outstanding capabilities in target detection, innovative methods have been introduced for infrared dim small target detection. This review comprehensively summarizes public datasets, the latest networks, and evaluation metrics for infrared dim small target detection. This review mainly focuses on deep learning methods from the past three years and categorizes them based on the six key issues in this field: (1) enhancing the representation capability of small targets; (2) improving the accuracy of bounding box regression; (3) resolving the issue of target information loss in the deep network; (4) balancing missed detections and false alarms; (5) adapting for complex backgrounds; (6) lightweight design and deployment issues of the network. Additionally, this review summarizes twelve public datasets for infrared dim small targets and evaluation metrics used for detection and quantitatively compares the performance of the latest networks. Finally, this review provides insights into the future directions of this field. In conclusion, this review aims to assist researchers in gaining a comprehensive understanding of the latest developments in infrared dim small target detection networks.

A Novel Approach to Detect Drones Using Deep Convolutional Neural Network Architecture.

Over the past decades, drones have become more attainable by the public due to their widespread availability at affordable prices. Nevertheless, this situation sparks serious concerns in both the cyber and physical security domains, as drones can be employed for malicious activities with public safety threats. However, detecting drones instantly and efficiently is a very difficult task due to their tiny size and swift flights. This paper presents a novel drone detection method using deep convolutional learning and deep transfer learning. The proposed algorithm employs a new feature extraction network, which is added to the modified YOU ONLY LOOK ONCE version2 (YOLOv2) network. The feature extraction model uses bypass connections to learn features from the training sets and solves the "vanishing gradient" problem caused by the increasing depth of the network. The structure of YOLOv2 is modified by replacing the rectified linear unit (relu) with a leaky-relu activation function and adding an extra convolutional layer with a stride of 2 to improve the small object detection accuracy. Using leaky-relu solves the "dying relu" problem. The additional convolution layer with a stride of 2 reduces the spatial dimensions of the feature maps and helps the network to focus on larger contextual information while still preserving the ability to detect small objects. The model is trained with a custom dataset that contains various types of drones, airplanes, birds, and helicopters under various weather conditions. The proposed model demonstrates a notable performance, achieving an accuracy of 77% on the test images with only 5 million learnable parameters in contrast to the Darknet53 + YOLOv3 model, which exhibits a 54% accuracy on the same test set despite employing 62 million learnable parameters.