Artificial Intelligence-based Methods Research Articles

Review of Short-Term Load Forecasting for Smart Grids Using Deep Neural Networks and Metaheuristic Methods

Forecasting electricity load demand is critical for power system planning and energy management. In particular, accurate short-term load forecasting (STLF), which focuses on the lead time horizon of few minutes to one week ahead, can help in better load scheduling, unit commitment, and cost-effective operation of smart power grids. In the last decade, different artificial intelligence (AI)-based techniques and metaheuristic algorithms have been utilized for STLF by the researchers and scientists with varying degrees of accuracy and efficacy. Despite the benefits of implemented methods for STLF, many drawbacks and associated problems have also been observed and reported by the researchers. This paper provides a comprehensive review of hybrid deep learning models based on nature-inspired metaheuristic techniques for STLF with respect to the analysis of the results and accuracy. Moreover, it also provides the research findings and gaps that will assist the researchers to have an early awareness of all important benefits and drawbacks of these integrated STLF methods scientifically and systematically. Especially, the hybrid forecast models using artificial intelligence-based methods for smart grids are focused. Several performance indices are used to compare and report the accuracy of these techniques including mean absolute percentage error (MAPE). Multiple other parametric and exogenous variable details have also been focused to figure out the potential of the intelligent load forecasting techniques from the perspective of smart power grids.

Characteristics of large-scale atmospheric circulation patterns conducive to severe spring and winter wind events over beijing in china based on a machine learning categorizing method

Severe wind events which occur in the metropolis of Beijing in China bring major catastrophes. Characteristics of severe winter and spring wind events over Beijing during the past 40 years have been analyzed. An artificial intelligence-based method is adopted to categorize the favorable large-scale circulation patterns and dominant weather systems. Four categories are concluded and compared to each other in terms of distributions of geopotential height at 500 hPa, temperature at 500 hPa, sea level pressure and their corresponding anomalies in 1979–2019. It is found that the first category (T1) which is dominated by strong cold trough at upper levels with strong cold-core high locating at surface is the most conducive circulation pattern, while the fourth category (T4) which is controlled by weak trough and strong ridge with strong low cyclone at surface is the least one. The second and third categories, represented by T2 and T3, are under the control of strong cold trough and warm ridge at upper levels with weak high at surface, and of weak trough and strong ridge with strong low cyclone at surface, respectively. Characteristics and differences under different backgrounds of global temperatures are analyzed by separating the past 40years into two distinct periods. The decreasing trends of intensities of the trough and ridge, the temperature at 500hPa, together with the surface systems, are found to be responsible for the decrease in severe wind events in T1, T2 and T3 in the last 20 years, while T4 is distinct to the other three categories with little change in its circulation pattern, and thus continues contributing to the severe wind events over Beijing. The results found in this study with the usage of an AI-based algorithm will benefit for the operational forecasting for extreme wind events over Beijing.

Artificial Intelligence Based Zero Trust Network

Model-based security metrics are an emerging topic of cyber security research that focuses on assessing an information system's risk exposure. We propose an end-to-end solution with the deployment of a zero-trust network utilising Artificial Intelligence in this article to understand the security posture of a system before it is rolled out and as it matures. The major part contains a discussion about the key methods and techniques which was utilized in the development process and simplified operation principles of each developed process. Some developed processes were tested practically to evaluate the problems in the processes. Modules for automatic processing and data analysis were also developed. These modules can be connected in case it is needed. The most important data collection methods were benchmarked to detect problematic situations in the operation in different realistic situations. With the perception from the benchmark test, the problematic parts of the data collection were discovered and proposals for the solution were made which could be developed and tested in the next iterations of the development process. Working Artificial intelligence-based detection and data enrichment methods were created. The results of the article allow multiple continuous research and development projects related to data collection and data analysis with statistical and artificial intelligence-based methods.

The diagnostic potential and barriers of microbiome based therapeutics.

High throughput technological innovations in the past decade have accelerated research into the trillions of commensal microbes in the gut. The 'omics' technologies used for microbiome analysis are constantly evolving, and large-scale datasets are being produced. Despite of the fact that much of the research is still in its early stages, specific microbial signatures have been associated with thepromotion of cancer, as well as other diseases such as inflammatory bowel disease, neurogenerative diareses etc. It has been also reported that the diversity of the gut microbiome influences the safety and efficacy of medicines. The availability and declining sequencing costs has rendered the employment of RNA-based diagnostics more common in the microbiome field necessitating improved data-analytical techniques so as to fully exploit all the resulting rich biological datasets, while accounting for their unique characteristics, such as their compositional nature as well their heterogeneity and sparsity. As a result, the gut microbiome is increasingly being demonstrating as an important component of personalised medicine since it not only plays a role in inter-individual variability in health and disease, but it also represents a potentially modifiable entity or feature that may be addressed by treatments in a personalised way. In this context, machine learning and artificial intelligence-based methods may be able to unveil new insights into biomedical analyses through the generation of models that may be used to predict category labels, and continuous values. Furthermore, diagnostic aspects will add value in the identification of the non invasive markers in the critical diseases like cancer.

DEVELOPMENT OF A GRAPHICAL USER INTERFACE TO SUPPORT THE SEMI-AUTOMATIC SEMANTIC SEGMENTATION OF UAS-IMAGES

Abstract. The development of remote sensing techniques dramatically improved the human knowledge of natural phenomena and the real time monitoring and interpretation of the events happening in the environment. The recently developed terrestrial, aerial and satellite remote sensors caused the availability of huge amount of data. The large size of such data is leading the research community to the search for efficient methods for real time information extraction, and, more in general, understanding the collected data. Nowadays, this is typically done by means of artificial intelligence-based methods, and, more specifically, usually by means of machine learning tools. Focusing on semantic segmentation, which is clearly related to a proper interpretation of the acquired remote sensing data, supervised machine learning is often used: it is based on the availability of a set of ground truth labeled data, which are used in order to properly train a machine learning classifier. Despite the latter, after a proper training phase, usually allows to obtain quite effective segmentation results, the ground truth labeled data production is usually a very laborious and time consuming task, performed by a human operator. Motivated by the latter consideration, this work aims at introducing a graphical interface developed in order to support semi-automatic semantic segmentation of images acquired by a UAS. Certain of the potentialities of the proposed graphical are shown in the specific case of plastic litter detection in multi-spectral images.

Improving accuracy of neuro fuzzy and support vector regression for drought modelling using grey wolf optimization

ABSTRACT In this study the capability of two advanced hybrid artificial intelligence-based methods is investigated in modelling meteorological droughts based on the standardized precipitation index (SPI) for various time windows. The outcomes are compared with adaptive neuro-fuzzy inference system (ANFIS), support vector regression (SVR), autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average (SARIMA). The best-fitted distribution functions are found to vary with respect to stations and time windows. For Canakkale, Istanbul and Tekirdag stations, the correlation coefficients (CC) of hybridized method of ANFIS with gray wolf optimization (ANFIS-GWO) and ARIMA are in the range of 0.88–0.94, 0.88–0.96 and 0.86–0.94, respectively. The performance index also showed that the ANFIS-GWO provides superior accuracy in modelling droughts, with the minimum value (PI = 0.78) for SPI12 of Canakkale station. Forward-chaining cross-validation and the P value of the Chi-squared test also confirm that ANFIS-GWO is the superior model, in which there is not a significant difference between the trend of the predicted categories and that of the real categories for all stations and SPIs.

Application of a Deep Learning Neural Network for Voiding Dysfunction Diagnosis Using a Vibration Sensor

In a clinical context, there are increasing numbers of people with voiding dysfunction. To date, the methods of monitoring the voiding status of patients have included voiding diary records at home or urodynamic examinations at hospitals. The former is less objective and often contains missing data, while the latter lacks frequent measurements and is an invasive procedure. In light of these shortcomings, this study developed an innovative and contact-free technique that assists in clinical voiding dysfunction monitoring and diagnosis. Vibration signals during urination were first detected using an accelerometer and then converted into the mel-frequency cepstrum coefficient (MFCC). Lastly, an artificial intelligence model combined with uniform manifold approximation and projection (UMAP) dimensionality reduction was used to analyze and predict six common patterns of uroflowmetry to assist in diagnosing voiding dysfunction. The model was applied to the voiding database, which included data from 76 males aged 30 to 80 who required uroflowmetry for voiding symptoms. The resulting system accuracy (precision, recall, and f1-score) was around 98% for both the weighted average and macro average. This low-cost system is suitable for at-home urinary monitoring and facilitates the long-term uroflow monitoring of patients outside hospital checkups. From a disease treatment and monitoring perspective, this article also reviews other studies and applications of artificial intelligence-based methods for voiding dysfunction monitoring, thus providing helpful diagnostic information for physicians.

Contexts and contradictions: a roadmap for computational drug repurposing with knowledge inference.

The cost of drug development continues to rise and may be prohibitive in cases of unmet clinical need, particularly for rare diseases. Artificial intelligence-based methods are promising in their potential to discover new treatment options. The task of drug repurposing hypothesis generation is well-posed as a link prediction problem in a knowledge graph (KG) of interacting of drugs, proteins, genes and disease phenotypes. KGs derived from biomedical literature are semantically rich and up-to-date representations of scientific knowledge. Inference methods on scientific KGs can be confounded by unspecified contexts and contradictions. Extracting context enables incorporation of relevant pharmacokinetic and pharmacodynamic detail, such as tissue specificity of interactions. Contradictions in biomedical KGs may arise when contexts are omitted or due to contradicting research claims. In this review, we describe challenges to creating literature-scale representations of pharmacological knowledge and survey current approaches toward incorporating context and resolving contradictions.

Computational Intelligence-Based Method for Automated Identification of COVID-19 and Pneumonia by Utilizing CXR Scans.

Chest X-ray (CXR) scans are emerging as an important diagnostic tool for the early spotting of COVID and other significant lung diseases. The recognition of visual symptoms is difficult and can take longer time by radiologists as CXR provides various signs of viral infection. Therefore, artificial intelligence-based method for automated identification of COVID by utilizing X-ray images has been found to be very promising. In the era of deep learning, effective utilization of existing pretrained generalized models is playing a decisive role in terms of time and accuracy. In this paper, the benefits of weights of existing pretrained model VGG16 and InceptionV3 have been taken. Base model has been created using pretrained models (VGG16 and InceptionV3). The last fully connected (FC) layer has been added as per the number of classes for classification of CXR in binary and multi-class classification by appropriately using transfer learning. Finally, combination of layers is made by integrating the FC layer weights of both the models (VGG16 and InceptionV3). The image dataset used for experimentation consists of healthy, COVID, pneumonia viral, and pneumonia bacterial. The proposed weight fusion method has outperformed the existing models in terms of accuracy, achieved 99.5% accuracy in binary classification over 20 epochs, and 98.2% accuracy in three-class classification over 100 epochs.

QR Code Detection

Due to the incorrect image collecting approach, QR code identification frequently confronts obstacles such as uneven backdrop fluctuations, inadequate illuminations, and distortions. As a result, identifying QR codes is difficult, and artificial intelligence-based methods were developed to address this challenge. This article uses an improved adaptive median filter technique and a QR code distortion correction method based on backpropagation (BP) neural networks to increase the identification rate of QR image codes. The distorted QR image can be fitted into the geometric deformation pattern using this combination of artificial intelligence methods, and QR code identification is possible. This paper addresses two-dimensional code distortion, which has been a critical research topic in present software systems. The research findings, which focused on the picture preprocessing step, demonstrated a significant improvement of 14 percent in the reading rate of QR image codes following processing using the system algorithm described in this article. The artificial intelligence technique used improves the recognition rate of the two-dimensional code image to some extent.

A New Artificial Intelligence-Based Method for Identifying Mycobacterium Tuberculosis in Ziehl-Neelsen Stain on Tissue.

Mycobacteria identification is crucial to diagnose tuberculosis. Since the bacillus is very small, finding it in Ziehl–Neelsen (ZN)-stained slides is a long task requiring significant pathologist’s effort. We developed an automated (AI-based) method of identification of mycobacteria. We prepared a training dataset of over 260,000 positive and over 700,000,000 negative patches annotated on scans of 510 whole slide images (WSI) of ZN-stained slides (110 positive and 400 negative). Several image augmentation techniques coupled with different custom computer vision architectures were used. WSIs automatic analysis was followed by a report indicating areas more likely to present mycobacteria. Our model performs AI-based diagnosis (the final decision of the diagnosis of WSI belongs to the pathologist). The results were validated internally on a dataset of 286,000 patches and tested in pathology laboratory settings on 60 ZN slides (23 positive and 37 negative). We compared the pathologists’ results obtained by separately evaluating slides and WSIs with the results given by a pathologist aided by automatic analysis of WSIs. Our architecture showed 0.977 area under the receiver operating characteristic curve. The clinical test presented 98.33% accuracy, 95.65% sensitivity, and 100% specificity for the AI-assisted method, outperforming any other AI-based proposed methods for AFB detection.

Prognostic impact of artificial intelligence-based volumetric quantification of the solid part of the tumor in clinical stage 0-I adenocarcinoma

IntroductionThe size of the solid part of a tumor, as measured using thin-section computed tomography, can help predict disease prognosis in patients with early-stage lung cancer. Although three-dimensional volumetric analysis may be more useful than two-dimensional evaluation, measuring the solid part of some lesions is difficult using this methods. We developed an artificial intelligence-based analysis software that can distinguish the solid and non-solid parts (ground-grass opacity). This software calculates the solid part volume in a totally automated and reproducible manner. The predictive performance of the artificial intelligence software was evaluated in terms of survival or recurrence-free survival. MethodsWe analyzed the high-resolution computed tomography images of the primary lesion in 772 consecutive patients with clinical stage 0-I adenocarcinoma. We performed automated measurement of the solid part volume using an artificial intelligence-based algorithm in collaboration with FUJIFILM Corporation. The solid part size, the solid part volume based on traditional three-dimensional volumetric analysis, and the solid part volume based on artificial intelligence were compared. ResultsHigher areas under the curve related to the solid part volume were provided by the artificial intelligence-based method (0.752) than by the solid part size (0.722) and traditional three-dimensional volumetric analysis-based method (0.723). Multivariate analysis demonstrated that the solid part volume based on artificial intelligence was independently correlated with overall survival (P = 0.019) and recurrence-free survival (P < 0.001). ConclusionThe solid part volume measured by artificial intelligence was superior to conventional methods in predicting the prognosis of clinical stage 0-I adenocarcinoma.

Performance analysis of a novel artificial intelligence-based computational method on published ex vivo drug sensitivity data to support targeted treatment decisions in precision oncology.

e13618 Background: Comprehensive molecular profiling is readily available for clinical practice. An extensive amount of published evidence provides information about the potential functional relevance of many driver genes and genetic alterations. But, due to the large number of driver genes and alterations, and the long-tail frequency distribution of these alterations and their possible combinations, very few single driver alterations have reached high-enough-level evidence alone to support clinical decisions. We hypothesized that by aggregating evidence following logical principles of reasoning by a computational system, this vast molecular information can be used to improve personalized treatment decisions. In the present study, we used previously published ex vivo drug sensitivity data to analyze the performance of an artificial intelligence-based computational method, the digital drug assignment (DDA), which has shown utility by improving treatment decisions in case of complex molecular profiles acquired in the SHIVA01 trial. Methods: We selected 111 cases with whole-genome sequencing (WGS) and ex vivo drug sensitivity data of a previously published acute myeloid leukemia study (Tyner et al, 2018). WGS variants were filtered for a preselected hematology-related panel of 446 genes and uploaded to a DDA-based software system to calculate the aggregated evidence level (AEL) values of associated molecularly targeted agents. DDA-predicted sensitivity (or resistance) was defined as AEL &gt; 0 (or AEL &lt; 0) in the presence of at least one actionable driver. Area under the curve (AUC) values were used for determining the ex vivo sensitivity or resistance of leukemia cells to 40 approved drugs and 53 developmental compounds. Results: The AUC values were significantly different in the drug-sensitive and -resistant groups forecasted by the DDA (167.1 and 205.5, respectively, p &lt; 0.0001) and differed significantly from the average AUC value (194.6). Overall, sensitivity was correctly predicted in 66% of compound-sample pairs (n = 671). 88 approved drugs had AEL value over 1000; of these 73% were effective according to the ex vivo results. While forecasted resistance was confirmed in 63% of the cases. With only the most sensitive/resistant 20% of cases considered from the ex vivo data, sensitivity was accurately predicted in 75% of approved compound-sample pairs (n = 173). 37 approved drugs had AEL value over 1000; of these 81% were confirmed as sensitive. Conclusions: The DDA-based computational reasoning has a promising performance in forecasting sensitivity and resistance to a broad spectrum of targeted agents based on molecular information. Therefore, it has the potential to automate, standardize and improve complex molecular profile-based targeted treatment decisions.

Real-world data on the clinical utility of a novel artificial intelligence-based computational method to support treatment decisions in gastrointestinal cancers.

e13616 Background: Several molecularly targeted agents (MTA) are registered in the same indication, even in the same line, to treat gastrointestinal (GI) cancers. Comprehensive molecular profiling is readily available for many patients, but only a few biomarkers are used to make clinical decisions between on-label MTAs. Most tumors harbor a combination of driver alterations potentially affecting the sensitivity to MTAs according to hundreds or thousands of evidence items that can explain the heterogeneous clinical response of individual patients to these therapies. We have developed an AI-based computational method, an automated, evidence-based reasoning framework, called digital drug assignment (DDA) to provide decision support based on the complete molecular profile for each patient. DDA has been shown to potentially improve treatment decisions in case of complex molecular profiles acquired in the SHIVA01 trial. In the present study, we analyzed the clinical utility of DDA in on-label treatment decisions in GI cancers. Methods: We analyzed the next-generation sequencing (NGS) data obtained using 50-591-gene panels of 239 patients with GI tract cancers from our everyday practice. Molecular profiles were uploaded to a DDA-based software system to calculate the aggregated evidence level (AEL) values of associated MTAs registered in the same indication of selected GI cancers. Decision support rate (DSR) was calculated as the percentage of patients where DDA could support clinical decisions based on the presence of genetic alterations associated with any of the available MTAs. Results: Targeted treatment options registered in the same indications and the DDA-based decision support rates (DSR) are shown in the table. For example, DDA could support the decision between cetuximab and bevacizumab based on NGS results in 47% of RAS/RAF wild-type colorectal cancer cases. Larger NGS panels yielded higher rates of decision support (44% vs 52% with gene panels under and over 300 genes, respectively). Conclusions: DDA-based computational reasoning can support molecular profile-based treatment decisions among different on-label MTAs in a clinically meaningful fraction of GI cancers.[Table: see text]

Generalized fault diagnosis method of transmission lines using transfer learning technique

Recent artificial intelligence-based methods have shown great promise in the use of neural networks for real-time detection of transmission line faults and estimation of their locations. The expansion of power systems including transmission lines with various lengths have made the fault detection, classification, and location estimation process more challenging. Transmission line datasets are stream data which are continuously collected by various sensors and hence, require generalized and fast fault diagnosis approaches. Newly collected datasets including voltages and currents for faulty and non-faulty situations might not have adequate and accurate labels that are useful to train neural networks. In this paper, a novel transfer learning framework based on a pre-trained LeNet-5 convolutional neural network is proposed. This method is able to diagnose faults for different transmission line lengths and impedances by transferring the knowledge from a source convolutional neural network to predict a dissimilar target dataset. By transferring this knowledge, faults from various transmission lines, even without sufficient data samples with labels, can be diagnosed faster and more efficiently than the existing methods. To prove the feasibility and effectiveness of this methodology, seven different datasets that include various lengths of transmission lines are used. The robustness of the proposed methodology against the generator voltage fluctuations, variations in fault locations, fault inception angle, fault resistance, and phase difference between the two generators are well studied to prove the reliability of this technique for fault diagnosis of transmission lines.

On the inverse identification methods for forming plasticity models using full-field measurements

The simulation of deep drawing processes and its quality is intrinsically dependent on the accuracy of the constitutive model in reproducing the mechanical behaviour of the sheet metal material. Today, the calibration of elastoplastic models – correspondent to the inverse identification of the material parameters – often uses full-field measurements, through Digital Image Correlation (DIC) techniques, to capture non-homogeneous strain fields and states, coupled with non-straightforward numerical inverse methodologies. In the last decade, new parameter identification methodologies, such as the Finite Element Model Updating (FEMU), the Constitutive Equation Gap (CEG) method, the Equilibrium Gap Method (EGM) and the Virtual Fields Method (VFM) have been developed and have proven to be effective for non-linear plasticity models. Nonetheless, the FEMU and the VFM have distinguished themselves from the others. More recently, supervised Machine Learning (ML) techniques have been also used as an inverse identification method. These artificial intelligence-based methods use large datasets of numerical tests to train an inverse model in which the input is the history of the strain field and loads during the test, and the output are directly the material parameters.The goal of this paper is to analyse, compare and discuss these inverse identification methods, with particular focus on the FEMU, VFM, and ML methodologies. A heterogeneous tensile-load test is considered to compare in detail the FEMU, VFM, and ML strategies.

Monosodium Glutamate-Induced Mouse Model With Unique Diabetic Retinal Neuropathy Features and Artificial Intelligence Techniques for Quantitative Evaluation.

ObjectiveTo establish an artificial intelligence-based method to quantitatively evaluate subtle pathological changes in retinal nerve cells and synapses in monosodium glutamate (MSG) mice and provide an effective animal model and technique for quantitative evaluation of retinal neurocytopathies.MethodsICR mice were subcutaneously injected with MSG to establish a model of metabolic syndrome. We then established a mouse model of type 1 diabetes, type 2 diabetes, and KKAy mouse model as control. The HE sections of the retina were visualized using an optical microscope. AI technology was used for quantitative evaluation of the retinal lesions in each group of rats. The surface area custom parameters of the retinal nerve fiber layer (RNFL), inner plexiform layer (IPL), inner nuclear layer (INL), and outer plexiform layer (OPL) were defined as SR, SIPL, SINL, and SOPL, respectively. Their heights were defined as HR, HIPL, HINL, and HOPL, and the number of ganglion cells was defined as A. Then, the attention-augmented fully convolutional Unet network was used to segment the retinal HE images, and AI technology to identify retinal neurocytopathies quantitatively.ResultsThe attention-augmented fully convolutional Unet network increased PA and IOU parameters for INL, OPL, RNFL, and ganglion cells and was superior in recognizing fine structures. A quantitative AI identification of the height of each layer of the retina showed that the heights of the IPL and INL of the MSG model were significantly less than those of the control groups; the retinas of the other diabetic models did not exhibit this pathological feature. The RNFLs of type 2 diabetes were thinner, and the characteristics of retinopathy were not obvious in the other animal models. The pathological changes seen on HE images were consistent with the results of the quantitative AI evaluation. Immunohistochemistry results showed that NMDAR2A, GluR2, and NRG1 were significantly downregulated in the retina of MSG mice.ConclusionsThe MSG retinopathy model is closely associated with neurotransmitter abnormalities and exhibits important characteristics of retinal neurodegeneration, making it suitable for studying retinal neurocytopathies. The AI recognition technology for retinal images established in the present study can be used for the quantitative and objective evaluation of drug efficacy.

Dew point pressure of gas condensates, modeling and a comprehensive review on literature data

The accurate and in-time prediction of gas condensates dew point pressure (PDew) is of great importance regarding the technical and economic points of view for fluid characterization, reservoir performance calculations, planning the development of gas condensate reservoirs, and design and optimization of production systems. Although the laboratory tests provide the most accurate and reliable results, it is an expensive and time-consuming process sometimes associated with some errors. Artificial intelligence-based methods have emerged as promising tools in different aspects of engineering. In this study, after a thorough analysis of the gas condensate data samples, the application of different intelligent modeling is investigated. A databank of 721 data samples is gathered, and different intelligent methods approaches are used for modeling. The results of three different data mining methods are combined using Committee Machine Intelligent Systems (CMIS) in an attempt to receive more accurate results. Three different methods of arithmetic, geometric, and harmonic approaches are utilized to develop the CMIS model. It was concluded that the harmonic CMIS yields the best predictions by average absolute relative deviation (AARD) and R2 values of 3.456% and 0.9702, respectively. This novel CMIS model could successfully outperform all the developed initial models. Additionally, a literature review showed that the proposed model could outperform the previously published models including artificial intelligence, equation of state, and correlation-based method considering both prediction accuracy and data coverage.

CTGNet: Automatic Analysis of Fetal Heart Rate from Cardiotocograph Using Artificial Intelligence

Abstract Objective: This study investigates the efficacy of analyzing fetal heart rate (FHR) signals based on Artificial Intelligence to obtain a baseline calculation and identify accelerations/decelerations in the FHR through electronic fetal monitoring during labor. Methods: A total of 43,888 cardiotocograph(CTG) records of female patients in labor from January 2012 to December 2020 were collected from the NanFang Hospital of Southern Medical University. After filtering the data, 2341 FHR records were used for the study. The ObVue fetal monitoring system, manufactured by Lian-Med Technology Co. Ltd., was used to monitor the FHR signals for these pregnant women from the beginning of the first stage of labor to the end of delivery. Two obstetric experts together annotated the FHR signals in the system to determine the baseline as well as accelerations/decelerations of the FHR. Our cardiotocograph network (CTGNet) as well as traditional methods were then used to automatically analyze the baseline and acceleration/deceleration of the FHR signals. The results of calculations were compared with the annotations provided by the obstetric experts, and ten-fold cross-validation was applied to evaluate them. The root-mean-square difference (RMSD) between the baselines, acceleration F-measure (Acc.F-measure), deceleration F-measure (Dec.F-measure), and the morphological analysis discordance index (MADI) were used as evaluation metrics. The data were analyzed by using a paired t-test. Results: The proposed CTGNet was superior to the best traditional method, proposed by Mantel, in terms of the RMSD.BL (1.7935 ± 0.8099 vs. 2.0293 ± 0.9267, t = −3.55 , P = 0.004), Acc.F-measure (86.8562 ± 10.9422 vs. 72.2367 ± 14.2096, t = 12.43, P &lt;0.001), Dec.F-measure (72.1038 ± 33.2592 vs. 58.5040 ± 38.0276, t = 4.10, P &lt;0.001), SI (34.8277±20.9595 vs. 54.8049 ± 25.0265, t = −9.39, P &lt;0.001), and MADI (3.1741 ± 1.9901 vs. 3.7289 ± 2.7253, t = −2.74, P = 0.012). The proposed CTGNet thus had significant advantages over the best traditional method on all evaluation metrics. Conclusion: The proposed Artificial Intelligence-based method CTGNet delivers good performance in terms of the automatic analysis of FHR based on cardiotocograph data. It promises to be a key component of smart obstetrics systems of the future.

IDF21-0386 DFU Track: Building a novel, Artificial Intelligence-based method to monitor wound healing

Background: While Diabetic Foot Ulcer (DFU) prevention and treatment have improved dramatically, assessment of DFU healing remains outdated, presenting a problem for research and clinical practice. Recent new mobile applications for standardizing DFU images have been reported, however, they are not automated and require large manual efforts. Aim: This study aims to define a reliable and sensitive metric by using machine learning to quantify wound severity, taking into account its size, shape, ischaemic and infection severity using a standard 2D photograph. This single metric could be used to identify and triage non-healing DFUs, as well as track wound healing in response to treatment. Method: Serial photographs of DFUs in 28 patients (90% T2DM, mean age 63, 83% male) with varying severity of DFU (72% UT 1A-C), were collected and analysed. Images from sequential clinic visits were automatically aligned. Wounds were automatically localized and segmented. Results: We have characterized DFUs and quantified changes in DFU characteristics across visits. 91% of baseline images were correctly identified and localized by the artificial intelligence (AI)-based wound localisation method. Discussion: If successful, this measure will facilitate targeting of appropriate therapies to individual patients, improving healing outcomes and therefore resulting in fewer amputations.