Poor fine motor performance is an important feature of children's developmental coordination disorder. To improve the diagnosis efficiency of developmental coordination disorder, computer vision-based evaluation methods have become a research hot topic. Most of the current methods are based on the evaluation of artificial features, while this paper proposes an automatic evaluation method of children's fine movements based on deep learning. By extracting human key points from videos, the method extracts the time series features of key points and uses deep learning neural networks to classify and predict children's fine movements. The experimental results show that the highest accuracy of this method is 81%, which provides an effective tool for the auxiliary diagnosis of children's developmental coordination disorder.

With the rapid development of artificial intelligence, especially deep learning technology, various new technologies and applications based on face images have emerged. Obstructive sleep apnea (OSA) is the disease with the highest morbidity and the most serious long-term harm among childhood sleep breathing disorders, and it is increasingly receiving common attention from families and society. Children with the disease have a special facial appearance and require early identification and treatment to prevent it. However, the current diagnostic methods have problems such as being invasive, time-consuming, and expensive. The purpose of this article is to use graph neural network technology based on face images to establish an OSA auxiliary diagnosis strategy for children to achieve OSA screening and analysis. Therefore, this article first takes the facial landmarks as the analysis object, divides the face into six key areas, and selects important landmarks in these regions. On this basis, to better consider the relationship between important landmarks, a global collaborative recognition strategy is proposed. By extracting the implicit relationship between landmarks, face graph structure data is established. Finally, the OSA-GNN model is established to achieve OSA screening and auxiliary analysis in children. Compared with other related studies, this strategy not only has a stronger representation and generalisation ability but can also carry out clinical applications better, providing doctors with diagnostic suggestions.

Abstract Graph Convolutional Networks (GCNs) have been widely used in skeleton‐based action recognition. Though significant performance has been achieved, it is still challenging to effectively model the complex dynamics of skeleton sequences. A novel position‐aware spatio‐temporal GCN for skeleton‐based action recognition is proposed, where the positional encoding is investigated to enhance the capacity of typical baselines for comprehending the dynamic characteristics of action sequence. Specifically, the authors’ method systematically investigates the temporal position encoding and spatial position embedding, in favour of explicitly capturing the sequence ordering information and the identity information of nodes that are used in graphs. Additionally, to alleviate the redundancy and over‐smoothing problems of typical GCNs, the authors’ method further investigates a subgraph mask, which gears to mine the prominent subgraph patterns over the underlying graph, letting the model be robust against the impaction of some irrelevant joints. Extensive experiments on three large‐scale datasets demonstrate that our model can achieve competitive results comparing to the previous state‐of‐art methods.

Gastroscopy is an important step in the diagnosis of early gastric cancer. However, because the morphological manifestations of early gastric cancer are not obvious, endoscopists need long-term specialized training and experience accumulation to correctly identify early cancer through magnification gastroscopy. In this paper, the data set of gastroscopy image is collected and enhanced, and target detection method is combined with gastroscopy image. The Mask R-CNN+BiFPN model was proposed to enhance the feature fusion and improve the detection effect of early gastric cancer lesions. Compared with Mask R-CNN, the improved Mask R-CNN model has better performance, with the sensitivity and specificity of 91.67% and 88.95% in accurately labeled gastroscopic datasets, respectively, showing a good segmentation effect for surface swelling lesions.

As the medical aesthetic market is growing rapidly in China, orthodontic treatment is becoming very common among the adolescent population. However, there are countless doctor-patient disputes due to treatment results that do not meet patients' expectations, so there is an urgent need for a method to predict treatment results. With the development of artificial intelligence technology, generative adversarial network has provided us with a new way of thinking. The purpose of this paper is to accurately predict the face of patients after orthodontic treatment by using generative adversarial network. Therefore, we designed an evaluation index to reflect the difference between the algorithm predicted image and the patient's real image. After that, we designed a network based on Encoder-Decoder architecture to transform the vectors in StyleGAN latent space. Finally, we carried out experiments to verify the effectiveness of the evaluation index design and the advantages of the algorithm.

OBJECTIVE: We aim to investigate the application and accuracy of artificial intelligence (AI) methods for automated medical literature screening for systematic reviews. MATERIALS AND METHODS: We systematically searched PubMed, Embase, and IEEE Xplore Digital Library to identify potentially relevant studies. We included studies in automated literature screening that reported study question, source of dataset, and developed algorithm models for literature screening. The literature screening results by human investigators were considered to be the reference standard. Quantitative synthesis of the accuracy was conducted using a bivariate model. RESULTS: Eighty-six studies were included in our systematic review and 17 studies were further included for meta-analysis. The combined recall, specificity, and precision were 0.928 [95% confidence interval (CI), 0.878-0.958], 0.647 (95% CI, 0.442-0.809), and 0.200 (95% CI, 0.135-0.287) when achieving maximized recall, but were 0.708 (95% CI, 0.570-0.816), 0.921 (95% CI, 0.824-0.967), and 0.461 (95% CI, 0.375-0.549) when achieving maximized precision in the AI models. No significant difference was found in recall among subgroup analyses including the algorithms, the number of screened literatures, and the fraction of included literatures. DISCUSSION AND CONCLUSION: This systematic review and meta-analysis study showed that the recall is more important than the specificity or precision in literature screening, and a recall over 0.95 should be prioritized. We recommend to report the effectiveness indices of automatic algorithms separately. At the current stage manual literature screening is still indispensable for medical systematic reviews.

In order to effectively implement Obstructive Sleep Apnea (OSA) pre-screening, an OSA pre-recognition model based on Resnet50 network was proposed. The frontal and lateral faces of 1000 patients with OSA were collected, and a series of preprocessing operations were performed to augment the image data, which were then fed into the Resnet50 network. The experimental results show that compared with the frontal face data, the classification accuracy obtained by using the side face data input into the network is 17.4% higher on average, and the frontal and lateral face data were indeed helpful in the pre - screening of obstructive sleep apnea. The grad-cam method shows that the focus areas of the model presented in this paper basically coincide with the face areas diagnosed in medical clinic.

Key frame extraction is very important for video data processing. This paper mainly studies how to extract the key frames of the video data when transforming the medical action video data into image data, so as to process and mine the medical action video data with image processing technology in the later period. Because of the difference of the doctor for some special action concerns, the existing algorithm of key frames can not extract the important frames selected by the doctors exactly, so this paper will review the traditional method to extract key frames, and use a new method of key frame extraction (multi-frame dimensionality-reduction difference method) to extract the key actions that clinical doctors pay attention to. Compared with the traditional method, this method can extract the frame image concerned by the doctor better.

Diagnosing infants who are small for gestational age (SGA) at early stages could help physicians to introduce interventions for SGA infants earlier. Machine learning (ML) is envisioned as a tool to identify SGA infants. However, ML has not been widely studied in this field. To develop effective SGA prediction models, we conducted four groups of experiments that considered basic ML methods, imbalanced data, feature selection and the time characteristics of variables, respectively. Infants with SGA data collected from 2010 to 2013 with gestational weeks between 24 and 42 were detected. Support vector machine (SVM), random forest (RF), logistic regression (LR) and Sparse LR models were trained on 10-fold cross validation. Precision and the area under the curve (AUC) of the receiver operator characteristic curve were evaluated. For each group, the performance of SVM and Sparse LR was similarly well. LR without any sparsity penalties performed worst, possibly caused by the overfitting problem. With the combination of handling imbalanced data and feature selection, the RF ensemble classifier performed best, which even obtained the highest AUC value (0.8547) with the help of expert knowledge. In other cases, RF performed worse than Sparse LR and SVM, possibly because of fully grown trees.

In order to combine the diagnosis and treatment guidelines for chronic obstructive pulmonary disease (COPD) with medical knowledge of electronic medical records, a semi-automatic construction method for knowledge maps of chronic obstructive pulmonary disease was proposed. First, the schema concept layer is designed in a top-down manner for the diagnosis and treatment of chronic obstructive pulmonary disease. Knowledge is removed from the data in the electronic medical records of the China-Japan Friendship Hospital. Knowledge extraction of unstructured data is based on the conditional random field (CRF) combined with the symptom dictionary method, through the design of the experiment to extract disease symptoms, verified the accuracy and effectiveness of the design, and stored the knowledge of chronic obstructive pulmonary disease to Neo4j.

In order to realize automatic recognition and extraction of entities in unstructured medical texts, a model combining language model conditional random field algorithm (CRF) and Bi-directional Long Short-term Memory networks (BiLSTM) is proposed. We crawled 804 drug specifications for treating asthma from the Internet, and then quantized the normalized field of drug specification word by a vector as the input of the neural network. Compared with the traditional machine learning algorithm CRF model, the system accuracy, recall and F1 value are improved by 6.18%, 5.2% and 4.87%. This model is applicable to extract named entity information from drug specification.

Gel is a post-operative cleaning material with antibacterial effect, which helps patients recover after surgery. It is more and more popular in surgery, but it is still controversial in use. This study collected the electronic medical records of patients in a hospital for nearly three years, using a combination of a variety of special selection methods to process data and using random forest, support vector machine, LightGBM and XGBoost and other machine learning methods to predict the suitability of patients. The results show that polysaccharide gel is not suitable for all people, whether to use it should consider different situations. This paper has studied the applicability of medical gels to patients, and established a predictability model to provide data support for the clinical application of this expensive medical material.

The fundus is part of the eye, and any partof the fundus is referred to as a fundus disease. In recentyears, the incidence of eye diseases has increased. Accordingto some statistics, cataracts are the most common cause ofblindness. Considering that the fundus image is one of themost important medical references that contributes to diagnosis, cataract classification and grading based on fundus images aresignificant. The fundus image analysis was used to simulate thework of the ophthalmologist's fundus image, and the cataractdetection and grading activities were examined. This is the useof machine learning to perform fundus image classification; inpractice, label samples tend to cause large losses during machinelearning, so the number of label samples is usually limited. For the damage effects of cataract disease, this paper proposesan improved semi-supervised learning method to acquire someadditional information from unlabeled cataract fundus imagesto improve the accuracy of the basic model for training onlythe marker images. In this proposed approach, we focus onstrategies for updating instance weights and combining severalbinary classifiers into one powerful multi-classifier. First, weadjust the original fundus image and enhance it with a histogramequalization method. Second, we extracted three image features(textures, wavelets, and sketches) from the enhanced fundusimage [1]. Third, we train a semi-supervised model on threeimage features. Finally, we combine several binary classifiersinto one powerful multi-classifier. The cataract fundus imagewill be divided into normal, mild, moderate and severe. Throughexperiments, the overall accuracy of the four categories on thetest data set is about 88.60%, 1.1% higher than Qiao's 87.52%, and 2.6% higher than the Song's 86.0% [2].

Magnetic resonance imaging (MRI) motion artifact is common in clinic which affects the doctor to accurately locate the lesion and diagnose the condition. MRI motion artifact is caused by the physiological movements of the patient while scanning the organ. Most of the current methods do artifact suppression and image restoration on the inverse Fourier transform level. They are neither effective nor efficient and can not be utilized in clinic. In this paper, the method that transfers deep learning into this domain with adopting a novel approach in Multi-scale mechanism for MRI motion artifact correction was proposed. What' more, a newer residual block with the deeper and wider architecture was proposed. With the deeper and wider residual block, the correction effect is greatly improved. The Peak Signal to Noise Ratio (PSNR) and Structural Similarity (SSIM) were adopted as the evaluation metrics. In short, our model is trainable in an end-to-end network, can be tested in real-time and achieves the state-of-the-art results for MRI motion artifact correction.

Retinal fundus image can perceive deep-seated blood vessels in the human body in a non-invasive manner. Retinal blood vessels are the primary anatomical structure that can be visible in the fundus image, while changes in the structural feature of retinal blood vessels cannot only reflect all sort of pathological changes but also serve as an important evidence for diagnosing cataract and other diseases. Automatic fundus image processing and analyzing in the computer has a significant effect on the auxiliary medical diagnosis. Moreover, the blood vessels extracted can be used as a feature for the classification of cataract fundus images. Most of the blood vessel extraction methods often used a heuristic feature set that are usually be extracted manually. For the limitations of current methods, we propose to use deep learning to identify blood vessels, which can perform automatic feature learning. We collected the dataset containing fundus images of 5620 patients for the extraction of blood vessels. We then performed Preprocessing by extracting green channel components and histogram equalization. We also present FCN structure in the fusion of dual sources in which preprocessed grayscale image and the edge information processed by the Sobel operators are used as an input. We also document that FCN enhance the richness of the input features and improve the accuracy. It can be concluded that the proposed method achieves the optimal accuracy for recognizing blood vessels of patients with cataract. Moreover, the accuracy of extracting normal fundus vessels reaches 94.91%. Furthermore, we are intended to use this proposed method for the vascular identification of other medical images.

Cataract is defined as a lenticular opacity presenting usually with poor visual acuity. It is considered the most common cause of blindness. Early diagnosis and treatment can reduce the suffering of patients and prevent visual impairment from turning into blindness. Recently, cataract diagnosis applying pattern recognition is in a rising period. For retinal fundus images, the task is usually cataract classification. However, it needs complex manual processing, which demands dexterous people and time taking exertion. Besides, it faces the challenge of effective interpretability and dependability. In this paper, we develop a deep-learning algorithm to intuitively identify cataract attributes to solve these limitations. Our model, is a 18(50)-layer convolutional neural network that inputs retinal image in G channel and outputs the prediction with heatmap. The heatmap localizes the areas where most indicative of different levels of cataract. Furthermore, we extend the training strategy for the corresponding task, which aims at improving the performance of the network. Comparing with other methods in cataract classification, we succeeded to achieve state of the art accuracy of proposed method on detection and grading task. Most importantly, our model provides a compelling reason via localizing the areas revealing cataract in the image.

Electronic medical records (EMRs) have high value for research, as they contain the patient's personal information, medical history, clinical examination, treatment process, and other information. Analysis based on EMRs can effectively assist doctors in clinical decision-making, provide data support for clinical research as well as personalized healthcare service for patients. We introduce a novel approach for EMR similarity computation by re-structuring and filtering some parts of physical examination result. Our approach is motivated by observations that it is easier to distinguish disease bias special part than bias the whole EMR which maybe contain some ineffective information. Assuming the check parts are independent, we split them and select effective parts. Then, we apply Deep NLP, converting the word to vectors which can be used to measure syntactic and semantic word similarities better. In addition, We replace traditional Euclidean distance with Word Mover's Distance(WMD), a novel distance function between text documents. Finally, KNN cluster is been used to evaluate the similarity between EMRs. Compared with traditional method such as LDA and LSI, our proposed method achieved higher recall value of disease classification problem.

Oral squamous cell carcinomas (OSCC) is the most common head and neck cancer worldwide, with more than 300,000 new cases being diagnosed annually. Studies have shown that miRNAs are involved in the process of growth, differentiation, apoptosis, invasion and metastasis of OSCC tumor cells. How miRNAs work together to contribute to this process is still largely unknown. The goal of our study was to characterize the coexpression network of miRNAs and to identify the miRNA subnetworks (modules) that were significantly associated with the OSCC cancer status. We also searched hub miRNAs that might play a vital role in the development of OSCC. We applied the weighted gene co-expression network analysis (WGCNA) to the miRNA expression profile data from a paired design study contributed by Shiah et al. To account for the within-pair correlation, a linear mixed model (LMM) was constructed to test the associations of miRNA modules to cancer status. Two significant modules (turquoise module with 254 miRNAs and grey module with 309 miRNAs) were identified. The miRNA miR-let-7c was the hub miRNA in the turquoise module in terms of node degree. Finally, we used miRsystem to perform the target gene prediction and KEGG pathway enrichment analysis of miRNAs within the two modules. Interestingly, the two modules have similar sets of target genes so that the top 6 enriched KEGG pathways for the 2 modules were the same. Compared with the probe-wise test used by Shiah et al., we took the network approach and identified significant OSCC-associated miRNA modules, which could help uncover the mechanism that miRNAs interplay each other to contribute to OSCC.

Medical search technologies are crucial to enable the user to rapidly and effectively discover useful information from massive medical and clinical data. Because of the complexity of medical terminology, traditional information search methods have not fully expressed the intention of the query request and explored the potential semantic knowledge in the document. In this paper, we propose a multi-analysis approach by considering the medical ontology as a semantic resource, which can excavate latent semantic information of a user's query request. In addition, we also recognize topics of medical documents to express text contents for providing support for calculating the similarity between query keywords and documents. Our experiments on PubMed medical article collections show that the semantic-based multi-analysis approach is feasible and efficient compared with other traditional approaches in medical retrieval.

The practice of using electronic medical records (EMR) to store healthcare data instead of traditional paper is becoming popular, as it's widely believed that the efficiency of medical institution could be improved and the cost could be reduced by using EMR. Nevertheless, EMR makes the sensitive healthcare data much easier to collect, process, store and publish. The change has increased the privacy risk of patients significantly. There are still many problems about patients' privacy protection in practical application, although many methods have been proposed to handle privacy risk when using EMR. Many hospitals are trying to make use of accumulated free text EMR for medical research and training. However, the hospitals may lose control with the EMR as the EMR may be transferred to different organizations in such application. And properly dealing with patients' privacy in EMR is an important requirement to avoid privacy leakage in such applications. In this paper, we study the privacy protection challenge those Hospitals are facing when they try to share EMR with other institutions. Then, we identify the major problems that make the traditional privacy preserving methods not suitable to process free text EMR, especially for free text Chinese EMR. Furthermore, we propose a new method to solve the problems and design a privacy-preserving data publishing system for electronic medical records based on it. Experiments with real-life data is showed to evaluate the new method.

As healthcare data is quite valuable to many organizations for scientific research or analysis, the demand of sharing healthcare data have been growing rapidly. Nevertheless, health care data usually contains a lot of patient privacy. Sharing that data directly would bring huge threaten to patient privacy. It's necessary to develop practical methods to balance health care data sharing and privacy protection. Although many approaches have been developed to deal with these problems, most of them are focusing on a small scope of the problem with single theory. In this paper, we'd like to introduce a framework for privacy preserving data sharing with the view of practical application in more comprehensive way. The framework focuses on three key problems of privacy protection during data sharing which are privacy definition and detection, privacy protection policy management, privacy preserving health care data sharing. And solutions to these three problems are discussed in details. A simple implementation of the framework would be introduced to solve the problems of privacy-preserving electronic medical records publishing.

We established a regional cooperative emergency care system of ST-elevation myocardial infarction patients based on the internet of things. In this article, the current status and problems of ST-elevation myocardial infarction patient emergency care have been studied and key influence factors are found. As the results, a shorter time from symptom onset to reperfusion is achieved with improved outcomes for patients with ST-segment elevation myocardial infarction (STEMI). Primary percutaneous coronary intervention (PCI) in patients with STEMI significantly reduces mortality and morbidity, particularly when door-to-balloon (D2B) time is <; 90 min. An expedited pre-hospital diagnosis and transfer pathway was developed, with rapid reperfusion times and favorable outcomes.

RFID has received considerable attention within the healthcare for almost a decade now. The technology's promise to efficiently track hospital supplies, medical equipment, medications and patients is an attractive proposition to the healthcare industry. However, the prospect of wide spread use of RFID tags in healthcare has also triggered discussions regarding privacy, particularly because RFID data in transit may easily be intercepted. In a nutshell, this technology has not really seen its true potential in healthcare since privacy concerns raised by the tag bearers are not properly addressed by existing protocols and frameworks. The two major types of privacy preservation techniques that are required in an RFID based healthcare are: 1) a privacy preserving authentication protocol is required while sensing RFID tags for different identification and monitoring purposes 2) a privacy preserving access control mechanism is required to restrict unauthorized access of private information while providing healthcare services using the tag ID. In this paper, we propose a component based framework (PriSens-HSAC) that makes an effort to address the above mentioned two privacy issues. To the best of our knowledge, this is the first framework to provide better privacy in RFID based healthcare systems, using authentication and access control technique.

With the improvement of living standards and the intensified social competition, the population of various chronic diseases is gradually expanded. In this study, we use a mature domestic commercial Body Area Network as our experiment platform-IVT mhealth system, a remote health care and rescue system created by IVT Corporation. This system adopts several most advanced technologies and patents in the world. It is a remote health care and rescue system, which consists of sensors, the call center and network platform. With the collaboration of medical institutions and health advisory center, it has the function of monitoring, positioning and asking for help. This system has achieved a dynamic measurement of ECG, blood pressure, blood glucose and blood oxygen. The data of users collected by wireless blood pressure meter, oximeter, ECG analyzer can be automatically sent to the cloud via cell phone and be kept as materials on records. This system provides a good platform for remote guidance, quantitative real-time monitoring and timely two-way feedback, as well as dynamic evaluation and overall adjustment for exercise rehabilitation of people on a large scale.