Search

The use of multiparametric magnetic resonance imaging (mpMRI)-derived radiomics has the potential to offer noninvasive, imaging-based biomarkers for the identification of subvisual characteristics indicative of a poor oncologic outcome. The present study, therefore, seeks to develop, validate, and assess the performance of an MRI-derived radiomic model for the prediction of prostate cancer (PC) recurrence following radical prostatectomy (RP) with curative intent. mpMRI imaging was obtained from 251 patients who had undergone an RP for the treatment of localized prostate cancer across two institutions and three surgeons. All patients had a minimum of 2 years follow-up via prostate-specific antigen serum testing. Each prostate mpMRI was individually reviewed, and the prostate was delineated as a single slice (ROI) on axial T2 high-resolution image sets. A total of 924 radiomic features were extracted and tested for stability via intraclass correlation coefficient (ICC) following image normalization via histogram matching. Fourteen important and nonredundant features were found to be predictors of PC recurrence at a mean ± SD of 3.2 ± 2.2 years post-RP. Five-fold, ten-run cross-validation of the model containing these fourteen features yielded an area under the curve (AUC) of 0.89 ± 0.04 in the training set (n = 225). In comparison, the University of California San Fransisco Cancer of the Prostate Risk Assessment score (UCSF-CAPRA) and Memorial Sloan Kettering Cancer Center (MSKCC) Pre-Radical prostatectomy nomograms yielded AUC of 0.66 ± 0.05 and 0.67 ± 0.05, respectively (p < 0.01). When the radiomic model was applied to the test set (n = 26), AUC was 0.78; sensitivity, specificity, positive predictive value, and negative predictive value were 60%, 86%, 52%, and 89%, respectively. Accuracy in predicting PC recurrence was 81%.

Purpose: While mammography has excellent sensitivity for the detection of breast lesions, its specificity is limited. Adjunct screening with ultrasound may partially alleviate this issue but also increases false positives, resulting in unnecessary biopsies. Our study investigated the use of Google AutoML Vision (Mountain View, California), a commercially available machine learning service, to both identify and characterize indeterminate breast lesions on ultrasound. Approach: B-mode images from 253 independent cases of indeterminate breast lesions scheduled for core biopsy were used for model creation and validation. The performances of two sub-models from AutoML Vision, the image classification model and object detection model, were evaluated, while also investigating training strategies to enhance model performances. Pathology from the patients biopsy was used as a reference standard. Results: The image classification models trained under different conditions demonstrated areas under the precision-recall curve (AUC) ranging from 0.85 to 0.96 during internal validation. Once deployed, the model with highest internal performance demonstrated a sensitivity of 100% [95% confidence interval (CI) of 73.5% to 100%], specificity of 83.3% (CI=51.6% to 97.9%), positive predictive value (PPV) of 85.7% (CI=62.9% to 95.5%), and negative predictive value (NPV) of 100% (CI non-evaluable) in an independent dataset. The object detection model demonstrated lower performance internally during development (AUC=0.67) and during prediction in the independent dataset [sensitivity=75% (CI=42.8 to 94.5), specificity=80% (CI=51.9 to 95.7), PPV=75% (CI=50.8 to 90.0), and NPV=80% (CI=59.3% to 91.7%)], but was able to demonstrate the location of the lesion within the image. Conclusions: Two models appear to be useful tools for identifying and classifying suspicious areas on B-mode images of indeterminate breast lesions.

Hexafluoropropylene Oxide Dimer Acid (HFPO-DA or GenX) is a pervasive perfluorinated compound with scant understood toxic effects. Toxicological studies on GenX have been conducted using animal models. To research deeper into the potential toxicity of GenX in humans and animals, we undertook a comprehensive analysis of transcriptome datasets across different species. A rank-in approach was utilized to merge different transcriptome datasets, and machine learning algorithms were employed to identify key genetic mechanisms common among various species and humans. We identified seven genes—TTR, ATP6V1B1, EPHX1, ITIH3, ATXN10, UBXN1, and HPX—as potential variables for classification of GenX-exposed samples, and the seven genes were verified in separate datasets of human, mouse, and rat samples. Bioinformatic analysis of the gene dataset further revealed that mitochondrial function and metabolic processes may be modulated by GenX through these key genes. Our findings provide insights into the underlying genetic mechanisms and toxicological impacts of GenX exposure across different species and offer valuable references for future studies using animal models to examine human exposure to GenX. [ABSTRACT FROM AUTHOR]

The simultaneous detection of multiple stimuli, such as pressure and temperature, has long been a persistent challenge for developing electronic skin (e‐skin) to emulate the functionality of human skin. Meanwhile, the demand for integrated power supply units is an additional pressing concern to achieve its lightweightness and flexibility. Herein, we propose a self‐powered dual temperature–pressure (SPDM) sensor, which utilizes a compressible ionic gel electrolyte driven by the potential difference between MXene and Al electrodes. The SPDM sensor exhibits a rapid and timely response to changes in pressure‐induced deformation, while exhibiting a slow and hysteretic response to temperature variations. These distinct response characteristics enable the differentiation of current signals generated by different stimuli through machine learning, resulting in an impressive accuracy rate of 99.1%. Furthermore, the developed SPDM sensor exhibits a wide pressure detection range of 0–800 kPa and a broad temperature detection range of 5–75°C, encompassing the environmental conditions encountered in daily human life. The dual‐mode coupled strategy by machine learning provides an effective approach for temperature and pressure detection and discrimination, showcasing its potential applications in wearable electronics, intelligent robots, human–machine interactions, and so on. [ABSTRACT FROM AUTHOR]

Handwriting difficulty is a defining feature of Chinese developmental dyslexia (DD) due to the complex structure and dense information contained within compound characters. Despite previous attempts to use deep neural network models to extract handwriting features, the temporal property of writing characters in sequential order during dictation tasks has been neglected. By combining transfer learning of convolutional neural network (CNN) and positional encoding with the temporal-sequential encoding of long short-term memory (LSTM) and attention mechanism, we trained and tested the model with handwriting images of 100,000 Chinese characters from 1064 children in Grades 2–6 (DD = 483; Typically Developing [TD] = 581). Using handwriting features only, the best model reached 83.2% accuracy, 79.2% sensitivity, 86.4% specificity, and 91.2% AUC. With grade information, the best model achieved 85.0% classification accuracy, 83.3% sensitivity, 86.4% specificity, and 89.7% AUC. These findings suggest the potential of utilizing machine learning technology to identify children at risk for dyslexia at an early age. [ABSTRACT FROM AUTHOR]

Traditional federated learning relies heavily on mature datasets, which typically consist of large volumes of uniformly distributed data. While acquiring extensive datasets is relatively straightforward in academic research, it becomes prohibitively expensive in practical applications, especially in emerging or specialized medical fields characterized by data scarcity. This poses a significant challenge. To address this issue, our study introduces a federated learning model that integrates few-shot learning techniques and is complemented by personalized knowledge distillation to further enhance the model's classification accuracy. This innovative approach significantly reduces the dependence on large-scale datasets, enabling efficient model training under limited data conditions. Our experimental evaluations conducted on small-scale datasets, including Omniglot, FC100, and mini-ImageNet, indicate that our model surpasses existing state-of-the-art federated learning models in terms of accuracy, achieving a substantial improvement. Specifically, on the FC100 dataset, the classification accuracy of the conventional federated learning algorithm FedAvg was merely 19.6%, whereas the method proposed in this study achieved a classification accuracy of 41%, representing an improvement of more than double. This advancement not only highlights our model's superiority in alleviating the challenges of limited data availability, but also expands the applicability of federated learning to a broader range of applications. [ABSTRACT FROM AUTHOR]

Malicious domains provide malware with covert communication channels which poses a severe threat to cybersecurity. Despite the continuous progress in detecting malicious domains with various machine learning algorithms, maintaining up‐to‐date various samples with fine‐labeled data for training is difficult. To handle these issues and improve the detection accuracy, a novel malicious domain detection method named MDND‐SS‐PO is proposed that combines semi‐supervised learning and parameter optimization. The contributions of the study are as follows. First, the method extracts the statistical features of the IP address, TTL value, the NXDomain record, and the domain name query characteristics to discriminate Domain‐Flux and Fast‐Flux domain names simultaneously. Second, an improved DBSCAN based on the neighborhood division is designed to cluster labeled data and unlabeled data with low time consumption. Then, based on the clustering hypothesis, unlabeled data is tagged with pseudo‐label according to the cluster results, which aims to train a supervised classifier effectively. Finally, Gaussian process regression is used to optimize parameter settings of the algorithm. And the Silhouette index and F1 score are introduced to evaluate the optimization results. Experimental results show that the proposed method achieved a precise detection performance of 0.885 when the ratio of labeled data is 5%. [ABSTRACT FROM AUTHOR]

Early and accurate diagnosis of ear deformities in newborns is crucial for an effective non-surgical correction treatment, since this commonly seen ear anomalies would affect aesthetics and cause mental problems if untreated. It is not easy even for experienced physicians to diagnose the auricular deformities of newborns and the classification of the sub-types, because of the rich bio-metric features embedded in the ear shape. Machine learning has already been introduced to analyze the auricular shape. However, there is little publicly available datasets of ear images from newborns. We released a dataset that contains quality-controlled photos of 3,852 ears from 1,926 newborns. The dataset also contains medical diagnosis of the ear shape, and the health data of each newborn and its mother. Our aim is to provide a freely accessible dataset, which would facilitate researches related with ear anatomies, such as the AI-aided detection and classification of auricular deformities and medical risk analysis. [ABSTRACT FROM AUTHOR]

The creep life prediction of austenitic heat-resistant steel is necessary to guarantee the safe operation of the high-temperature components in thermal power plants. This work presents a machine learning model that can be applied to predict the creep life of austenitic steels, offering a novel method and approach for such predictions. In this paper, creep life data from six typical austenitic heat-resistant steels are used to predict their creep life using various machine learning models. Moreover, the dissimilarities between the machine learning model and the conventional lifetime prediction method are compared. Finally, the influence of different input characteristics on creep life is discussed. The results demonstrate that the prediction accuracy of machine learning depends on both the model and the dataset used. The Gaussian model based on the second dataset achieves the highest level of prediction accuracy. Additionally, the accuracy and the generalization ability of the machine learning model prediction are significantly better than those of the traditional model. Lastly, the effect of the input characteristics on creep life is generally consistent with experimental observations and theoretical analyses. [ABSTRACT FROM AUTHOR]

Seasonal freezing depth change is important in many environmental science and engineering applications. However, such changes are rare at region scales, especially over China, in the long time series. In this study, we evaluated the annual changes in seasonal maximum freezing depth (MFD) over China from 1971 to 2020 using an ensemble-modeling method based on support vector machine regression (SVMR) with 600 repetitions. Remote sensing data and climate data were input variables used as predictors. The models were trained using a large amount of annual measurement data from 600 meteorological stations. The main reason for using SVMR here was because it has been shown to perform better than random forests (RF), k-nearest neighbors (KNN), and generalized linear regression (GLR) in these cases. The prediction results were generally consistent with the observed MFD values. Cross validation showed that the model performed well on training data and had a better spatial generalization ability. The results show that the freezing depth of seasonally frozen ground in China decreased year by year. The average MFD was reduced by 3.64 cm, 7.59 cm, 5.54 cm, and 5.58 cm, in the 1980s, 1990s, 2000s, and 2010s, respectively, compared with the decade before. In the last 50 years, the area occupied by the freezing depth ranges of 0–40 cm, 40–60 cm, 60–80 cm, 80–100 cm, and 120–140 cm increased by 99,300 square kilometers, 146,200 square kilometers, 130,300 square kilometers, 115,600 square kilometers, and 83,800 square kilometers, respectively. In addition to the slight decrease in freezing depth range of 100–120 cm, the reduced area was 29,500 square kilometers. Freezing depth ranges greater than 140 cm showed a decreasing trend. The freezing depth range of 140–160 cm, which was the lowest range, decreased by 89,700 square kilometers. The 160–180 cm range decreased by 120,500 square kilometers, and the 180–200 cm range decreased by 161,500 square kilometers. The freezing depth range greater than 200 cm, which was the highest reduction range, decreased by 174,000 square kilometers. Considering the lack of data on the change in MFD of seasonally frozen ground in China in recent decades, machine learning provides an effective method for studying meteorological data and reanalysis data in order to predict the changes in MFD. [ABSTRACT FROM AUTHOR]

Based on the monthly precipitation data of 139 meteorological stations from 1981 to 2015, the trend line method was used to analyze the variation of the drought area and sown area from 1981 to 2015 and the characteristics of drought in Hebei Province. The results showed the following. (1) the frequency of drought was the highest in winter and the lowest in summer, which was 43.2 and 26.2%, respectively. (2) The ratio of drought stations was highest in winter and lowest in summer. The frequency of global drought was 37 and 17% in summer and winter, respectively. Except for the increasing trend of drought in summer, all other seasons showed a decreasing trend which was significant in autumn (α = 0.05). (3) The drought intensity of the four seasons was all less than 1. The drought intensity in spring and summer showed an upward trend, while that in autumn and winter showed a downward trend. The drought intensity in autumn showed a significant downward trend (α = 0.05). (4) Drought disaster rate was positively correlated with summer drought intensity and drought ratio. [ABSTRACT FROM AUTHOR]

This paper proposes an advanced expert collaborative filtering recommendation algorithm. Although ordinary expert system filtering algorithms have improved the recommendation accuracy of collaborative filtering technology to a certain extent, they have not screened the level of expertise of experts, and the credibility of experts varies. Therefore, this paper proposes an expert mining system based on signal fluctuations. The algorithm uses signal processing technology to filter the level of experts. This method introduces a kurtosis factor. Regarding the user's rating sequence as a random discrete signal, and then randomly sorting the user's ratings k times, the average kurtosis of the user is obtained. And take the kurtosis value as the credibility of expert users. Through experiments on multiple datasets including MovieLens, Jester, Booking-Crossings, and Last.fm, we have proved the advancement and reliability of our method. [ABSTRACT FROM AUTHOR]

This research suggests a multi-level federated edge learning algorithm by leveraging the advantages of Edge Computing Paradigm. Model aggregation is partially moved from a cloud center server to edge servers in this framework, and edge servers are connected hierarchically depending on where they are located and how much computational power they have. At the same time, we considered an important issue: the heterogeneity of different client computing resources (such as device processor computing power) and server communication channels (which may be limited by geography or device). For this situation, a client and edge server selection algorithm ( C E S A ) based on a greedy algorithm is proposed in this paper. Given resource constraints, C E S A aims to select as many clients and edge servers as possible to participate in the model computation in order to improve the accuracy of the model. The simulation results show that, when the number of clients is high, the multi-level federated edge learning algorithm can shorten the model training time and improve efficiency compared to the traditional federated learning algorithm. Meanwhile, the C E S A is able to aggregate more clients for training in the same amount of time compared to the baseline algorithm, improving model training accuracy. [ABSTRACT FROM AUTHOR]

In construction planning, decision making has a great impact on final project performance. Hence, it is essential for project managers to assess the construction planning and make informed decisions. However, disproportionately large uncertainties occur during the construction planning stage; in the worst case, reliable probability distributions of uncertainties are sometimes unavailable due to a lack of information before construction implementation. This situation constitutes a deep uncertainty problem, making it a challenge to perform a probability-based uncertainty assessment. The current study proposes a modeling approach that applies prediction intervals for construction planning via the integration of discrete-event simulation (DES), fuzzy C-means clustering (FCM), Bayesian regularization backpropagation neural networks (BRBNNs), and particle swarm optimization (PSO). The DES is used to perform data sampling of the construction alternatives and assess their performances under uncertainty. Based on the generated samples, the FCM, BRBNN, and PSO are integrated in a machine learning algorithm to model the prediction intervals that represent relationships between construction planning schemes, performances, and the corresponding uncertainties. The proposed approach was applied to a case project, with the results indicating that it is capable of modeling construction performance and deep uncertainties with a defined 95% confidence level and fluctuation within 1~9%. The presented research contributes a new and innovative option, using prediction intervals to solve deep uncertainty problems, without relying on the probability of the uncertainty. This study demonstrates the effectiveness of the proposed approach in construction planning. [ABSTRACT FROM AUTHOR]

Background: Acute pancreatitis (AP) with critical illness is linked to increased morbidity and mortality. Current risk scores to identify high-risk AP patients have certain limitations.Objective: To develop and validate a machine learning tool within 48 h after admission for predicting which patients with AP will develop critical illness based on ubiquitously available clinical, laboratory, and radiologic variables.Methods: 5460 AP patients were enrolled. Clinical, laboratory, and imaging variables were collected within 48 h after hospital admission. Least Absolute Shrinkage Selection Operator with bootstrap method was employed to select the most informative variables. Five different machine learning models were constructed to predictive likelihood of critical illness, and the optimal model (APCU) was selected. External cohort was used to validate APCU. APCU and other risk scores were compared using multivariate analysis. Models were evaluated by area under the curve (AUC). The decision curve analysis was employed to evaluate the standardized net benefit.Results: Xgboost was constructed and selected as APCU, involving age, comorbid disease, mental status, pulmonary infiltrates, procalcitonin (PCT), neutrophil percentage (Neu%), ALT/AST, ratio of albumin and globulin, cholinesterase, Urea, Glu, AST and serum total cholesterol. The APCU performed excellently in discriminating AP risk in internal cohort (AUC = 0.95) and external cohort (AUC = 0.873). The APCU was significant for biliogenic AP (OR = 4.25 [2.08-8.72], P < 0.001), alcoholic AP (OR = 3.60 [1.67-7.72], P = 0.001), hyperlipidemic AP (OR = 2.63 [1.28-5.37], P = 0.008) and tumor AP (OR = 4.57 [2.14-9.72], P < 0.001). APCU yielded the highest clinical net benefit, comparatively.Conclusion: Machine learning tool based on ubiquitously available clinical variables accurately predicts the development of AP, optimizing the management of AP. [ABSTRACT FROM AUTHOR]

Attention mechanisms have significantly boosted the performance of video classification neural networks thanks to the utilization of perspective contexts. However, the current research on video attention generally focuses on adopting a specific aspect of contexts (e.g., channel, spatial/temporal, or global context) to refine the features and neglects their underlying correlation when computing attentions. This leads to incomplete context utilization and hence bears the weakness of limited performance improvement. To tackle the problem, this paper proposes an efficient attention-in-attention (AIA) method for element-wise feature refinement, which investigates the feasibility of inserting the channel context into the spatio-temporal attention learning module, referred to as CinST, and also its reverse variant, referred to as STinC. Specifically, we instantiate the video feature contexts as dynamics aggregated along a specific axis with global average and max pooling operations. The workflow of an AIA module is that the first attention block uses one kind of context information to guide the gating weights calculation of the second attention that targets at the other context. Moreover, all the computational operations in attention units act on the pooled dimension, which results in quite few computational cost increase (< 0.02%). To verify our method, we densely integrate it into two classical video network backbones and conduct extensive experiments on several standard video classification benchmarks. The source code of our AIA is available at https://github.com/haoyanbin918/Attention-in-Attention. [ABSTRACT FROM AUTHOR]

The aim of this study was to determine whether left ventricular ejection fraction (LVEF) and right ventricular ejection fraction (RVEF) and left ventricular mass (LVM) measurements made using 3 fully automated deep learning (DL) algorithms are accurate and interchangeable and can be used to classify ventricular function and risk-stratify patients as accurately as an expert. Artificial intelligence is increasingly used to assess cardiac function and LVM from cardiac magnetic resonance images. Two hundred patients were identified from a registry of individuals who underwent vasodilator stress cardiac magnetic resonance. LVEF, LVM, and RVEF were determined using 3 fully automated commercial DL algorithms and by a clinical expert (CLIN) using conventional methodology. Additionally, LVEF values were classified according to clinically important ranges: <35%, 35% to 50%, and ≥50%. Both ejection fraction values and classifications made by the DL ejection fraction approaches were compared against CLIN ejection fraction reference. Receiver-operating characteristic curve analysis was performed to evaluate the ability of CLIN and each of the DL classifications to predict major adverse cardiovascular events. Excellent correlations were seen for each DL-LVEF compared with CLIN-LVEF (r = 0.83-0.93). Good correlations were present between DL-LVM and CLIN-LVM (r = 0.75-0.85). Modest correlations were observed between DL-RVEF and CLIN-RVEF (r = 0.59-0.68). A >10% error between CLIN and DL ejection fraction was present in 5% to 18% of cases for the left ventricle and 23% to 43% for the right ventricle. LVEF classification agreed with CLIN-LVEF classification in 86%, 80%, and 85% cases for the 3 DL-LVEF approaches. There were no differences among the 4 approaches in associations with major adverse cardiovascular events for LVEF, LVM, and RVEF. This study revealed good agreement between automated and expert-derived LVEF and similarly strong associations with outcomes, compared with an expert. However, the ability of these automated measurements to accurately classify left ventricular function for treatment decision remains limited. DL-LVM showed good agreement with CLIN-LVM. DL-RVEF approaches need further refinements. [Display omitted] [ABSTRACT FROM AUTHOR]

Objectives: To investigate the association between radiomics features and epilepsy in patients with unruptured brain arteriovenous malformations (bAVMs) and to develop a prediction model based on radiomics features and clinical characteristics for bAVM-related epilepsy. Methods: This retrospective study enrolled 176 patients with unruptured bAVMs. After manual lesion segmentation, a total of 858 radiomics features were extracted from time-of-flight magnetic resonance angiography (TOF-MRA). A radiomics model was constructed, and a radiomics score was calculated. Meanwhile, the demographic and angioarchitectural characteristics of patients were assessed to build a clinical model. Incorporating the radiomics score and independent clinical risk factors, a combined model was constructed. The performance of the models was assessed with respect to discrimination, calibration, and clinical usefulness. Results: The clinical model incorporating 3 clinical features had an area under the curve (AUC) of 0.71. Fifteen radiomics features were used to build the radiomics model, which had a higher AUC of 0.78. Incorporating the radiomics score and clinical risk factors, the combined model showed a favorable discrimination ability and calibration, with an AUC of 0.82. Decision curve analysis (DCA) demonstrated that the combined model outperformed the clinical model and radiomics model in terms of clinical usefulness. Conclusions: The radiomics features extracted from TOF-MRA were associated with epilepsy in patients with unruptured bAVMs. The radiomics-clinical nomogram, which was constructed based on the model incorporating the radiomics score and clinical features, showed favorable predictive efficacy for bAVM-related epilepsy. [ABSTRACT FROM AUTHOR]

Previous deep learning methods usually required large‐scale annotated data, which is computationally exhaustive and unrealistic in certain scenarios. Therefore, few‐shot learning, where only a few annotated training images are available for training, has attracted increasing attention these days, showing huge potential in practical applications, such as portable equipment or security inspection, and so on. However, current few‐shot learning methods usually neglect the valuable semantic correlations between samples, thereby failing in extracting discriminating relations to achieve accurate predictive results. In this work, extending on a recent state‐of‐the‐art few‐shot learning method, transductive relation‐propagation network (TRPN), which considers the correlations between training samples, a constrained relation‐propagation network is proposed to further regularise the distilled correlations and thus achieve favourable few‐shot classification performance. The proposed framework contains three main components, namely preprocess module, relational propagation module, and relation constraint module. First, sample features are extracted and a relation graph node is constructed by treating the relation of each support–query pair as a graph node in the preprocess module. After that, in the relation propagation module (RPM), the valuable information of support–query pairs is modelled and propagated to directly generate the relational representations for further prediction. Then, a relation constraint module is introduced to regularise the relational representations and make it consistent with the ground‐truth relations as much as possible. With the guidance of the effective RPM and relation constraint module, the relational representations of the support–query pairs are distinguishable and thus can achieve accurate predictive results. Comprehensive experiments conducted on widely used benchmarks validate the effectiveness of our method compared to state‐of‐the‐art few‐shot classification approaches. [ABSTRACT FROM AUTHOR]

Background: The role of next generation sequencing (NGS) for identifying high risk mutations in thyroid nodules following fine needle aspiration (FNA) biopsy continues to grow. However, ultrasound diagnosis even using the American College of Radiology's Thyroid Imaging Reporting and Data System (TI-RADS) has limited ability to stratify genetic risk. The purpose of this study was to incorporate an artificial intelligence (AI) algorithm of thyroid ultrasound with object detection within the TI-RADS scoring system to improve prediction of genetic risk in these nodules. Methods: Two hundred fifty-two nodules from 249 patients that underwent ultrasound imaging and ultrasound-guided FNA with NGS with or without resection were retrospectively selected for this study. A machine learning program (Google AutoML) was employed for both automated nodule identification and risk stratification. Two hundred one nodules were used for model training and 51 reserved for testing. Three blinded radiologists scored the images of the test set nodules using TI-RADS and assigned each nodule as high or low risk based on the presence of highly suspicious imaging features on TI-RADS (very hypoechoic, taller-than-wide, extra-thyroidal extension, punctate echogenic foci). Subsequently, the TI-RADS classification was modified to incorporate AI for T4 nodules while treating T1-3 as low risk and T5 as high risk. All diagnostic predictions were compared to the presence of a high-risk mutation and pathology when available. Results: The AI algorithm correctly located all nodules in the test dataset (100% object detection). The model predicted the malignancy risk with a sensitivity of 73.9%, specificity of 70.8%, positive predictive value (PPV) of 70.8%, negative predictive value (NPV) of 73.9% and accuracy of 72.4% during the testing. The radiologists performed with a sensitivity of 52.1 ± 4.4%, specificity of 65.2 ± 6.4%, PPV of 59.1 ± 3.5%, NPV of 58.7 ± 1.8%, and accuracy of 58.8 ± 2.5% when using TI-RADS and sensitivity of 53.6 ± 17.6% (p=0.87), specificity of 83.3 ± 7.2% (p=0.06), PPV of 75.7 ± 8.5% (p=0.13), NPV of 66.0 ± 8.8% (p=0.31), and accuracy of 68.7 ± 7.4% (p=0.21) when using AI-modified TI-RADS. Conclusions: Incorporation of AI into TI-RADS improved radiologist performance and showed better malignancy risk prediction than AI alone when classifying thyroid nodules. Employing AI in existing thyroid nodule classification systems may help more accurately identifying high-risk nodules. [ABSTRACT FROM AUTHOR]

Recent studies in brain science and neurological medicine paid a particular attention to develop machine learning-based techniques for the detection and prediction of epileptic seizures with electroencephalogram (EEG). As a noninvasive monitoring method to record brain electrical activities, EEG has been widely used for capturing the underlying dynamics of disruptive neuronal responses across the brain in real-time to provide clinical guidance in support of epileptic seizure treatments in practice. In this study, we introduce a novel dynamic learning method that first infers a time-varying network constituted by multivariate EEG signals, which represents the overall dynamics of the brain network, and subsequently quantifies its topological property using graph theory. We demonstrate the efficacy of our learning method to detect relatively strong synchronization (characterized by the algebraic connectivity metric) caused by abnormal neuronal firing during a seizure onset. The computational results for a realistic scalp EEG database show a detection rate of 93.6% and a false positive rate of 0.16 per hour (FP/h); furthermore, our method observes potential pre-seizure phenomena in some cases. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) tools have been applied to diagnose or predict disease risk from medical images with recent data disclosure actions, but few of them are designed for mobile terminals due to the limited computational power and storage capacity of mobile devices. In this work, a novel AI diagnostic system is proposed for cholelithiasis recognition on mobile devices with Android platform. To this aim, a data set of CT images of cholelithiasis is firstly collected from The Third Hospital of Shandong Province, China, and then we technically use histogram equalization to preprocess these CT images. As results, a lightweight convolutional neural network is obtained in a constructive way to extract cholelith features and recognize gallstones. In terms of implementation, we compile Java and C++ to adapt to the application of deep learning algorithm on mobile devices with Android platform. Noted that, the training task is completed offline on PC, but cholelithiasis recognition tasks are performed on mobile terminals. We evaluate and compare the performance of our MobileNetV2 with MobileNetV1, Single Shot Detector (SSD), YOLOv2 and original SSD (with VGG-16) as feature extractors for object detection. It is achieved that our MobileNetV2 achieve similar accuracy rate, about 91% with the other four methods, but the number of parameters used is reduced from 36.1M (SSD 300, SSD512), 50.7M (Yolov2) and 5.1M (MobileNetV1) to 4.3M (MobileNetV2). The complete process on testing mobile devices, including Virtual machine, Xiaomi 7 and Htc One M8 can be controlled within 4 seconds in recognizing cholelithiasis as well as the degree of the disease. [ABSTRACT FROM AUTHOR]

Locating diseases precisely from medical images, like ultrasonic and CT images, have been one of the most challenging problems in medical image analysis. In recent years, the vigorous development of deep learning models have greatly improved the accuracy in disease location on medical images. However, there are few artificial intelligent methods for identifying cholelithiasis and classifying gallstones on CT images, since no open source CT images dataset of cholelithiasis and gallstones is available for training the models and verifying their performance. In this paper, we build up the first medical image dataset of cholelithiasis by collecting 223846 CT images with gallstone of 1369 patients. With these CT images, a neural network is trained to “pick up” CT images of high quality as training set, and then a novel Yolo neural network, named Yolov3-arch neural network, is proposed to identify cholelithiasis and classify gallstones on CT images. Identification and classification accuracies are obtained by 10-fold cross-validations. It is obtained that our Yolov3-arch model is with average accuracy 92.7% in identifying granular gallstones and average accuracy 80.3% in identifying muddy gallstones. This achieves 3.5% and 8% improvements in identifying granular and muddy gallstones to general Yolo v3 model, respectively. Also, the average cholelithiasis identifying accuracy is improved to 86.50% from 80.75%. Meanwhile, our method can reduce the misdiagnosis rate of negative samples by the object detection model. [ABSTRACT FROM AUTHOR]

• Computing on edge end improves the efficiency and reliability of data analysis. • A lightweight CNN model is efficient for real-time crowd density estimation on edge. • Better crowd density inference speed with a slight increase in estimation accuracy. • Equip the model in an IoT device to monitor the crowd density in a Subway Station. Crowd stampedes and incidents are critical threats to public security that have caused countless deaths during the past few decades. To avoid crowd stampedes, real-time crowd density estimation can help monitor crowd movements, and thus support a timely evacuation strategy development. In previous studies, scholars and engineers developed multiple video-based crowd density estimation algorithms based on deep neural networks. The excessive computational complexity of deep learning algorithms exacerbated the algorithm's efficiency, causing unacceptable real-time performance. In the Internet of Things era, deploying the crowd density estimation task with edge computing is an advanced strategy to maintain the real-time performance of the entire system. Considering the limited computational resources on the edge devices, deep learning-based crowd density estimation algorithms normally cannot be handled. To fulfill the deployment on the edge device, the algorithms need to be optimized with a smaller model size. Therefore, this paper proposes a lightweight Convolutional Neural Networks (CNN) based crowd density estimation model by combining the modified MobileNetv2 and the dilated convolution. Public crowd image data sets are used to conduct experiments for evaluating the performance of the proposed algorithm in terms of accuracy and inference speed. The results show that our model achieves much better inference speed accompanied by a slight increase in accuracy. The proposed method of this study can enhance the performance of the crowd monitoring system, and therefore help avoid crowd stampedes and incidents. [ABSTRACT FROM AUTHOR]

Background: A challenge preventing routine clinical implementation of Monte Carlo (MC)-based lung SBRT is the difficulty of reinterpreting historical outcome data calculated with inaccurate dose algorithms, because the target dose was found to decrease to varying degrees when recalculated with MC. The large variability was previously found to be affected by factors such as tumour size, location, and lung density, usually through sub-group comparisons. We hereby conducted a pilot study to systematically and quantitatively analyze these patient factors and explore accurate target dose conversion models, so that large-scale historical outcome data can be correlated with more accurate MC dose without recalculation.Methods: Twenty-one patients that underwent SBRT for early-stage lung cancer were replanned with 6MV 360° dynamic conformal arcs using pencil-beam (PB) and recalculated with MC. The percent D95 difference (PB-MC) was calculated for the PTV and GTV. Using single linear regression, this difference was correlated with the following quantitative patient indices: maximum tumour diameter (MaxD); PTV and GTV volumes; minimum distance from tumour to soft tissue (dmin); and mean density and standard deviation of the PTV, GTV, PTV margin, lung, and 2 mm, 15 mm, 50 mm shells outside the PTV. Multiple linear regression and artificial neural network (ANN) were employed to model multiple factors and improve dose conversion accuracy.Results: Single linear regression with PTV D95 deficiency identified the strongest correlation on mean-density (location) indices, weaker on lung density, and the weakest on size indices, with the following R(2) values in decreasing orders: shell2mm (0.71), PTV (0.68), PTV margin (0.65), shell15mm (0.62), shell50mm (0.49), lung (0.40), dmin (0.22), GTV (0.19), MaxD (0.17), PTV volume (0.15), and GTV volume (0.08). A multiple linear regression model yielded the significance factor of 3.0E-7 using two independent features: mean density of shell2mm (P = 1.6E-7) and PTV volume (P = 0.006). A 4-feature ANN model slightly improved the modeling accuracy.Conclusion: Quantifiable density features were proposed, replacing simple central/peripheral location designation, which showed strong correlations with PB-to-MC target dose conversion magnitude, followed by lung density and target size. Density in the immediate outer and inner areas of the PTV showed the strongest correlations. A multiple linear regression model with one such feature and PTV volume established a high significance factor, improving dose conversion accuracy. [ABSTRACT FROM AUTHOR]

Class evolution, the phenomenon of class emergence and disappearance, is an important research topic for data stream mining. All previous studies implicitly regard class evolution as a transient change, which is not true for many real-world problems. This paper concerns the scenario where classes emerge or disappear gradually. A class-based ensemble approach, namely Class-Based ensemble for Class Evolution (CBCE), is proposed. By maintaining a base learner for each class and dynamically updating the base learners with new data, CBCE can rapidly adjust to class evolution. A novel under-sampling method for the base learners is also proposed to handle the dynamic class-imbalance problem caused by the gradual evolution of classes. Empirical studies demonstrate the effectiveness of CBCE in various class evolution scenarios in comparison to existing class evolution adaptation methods. [ABSTRACT FROM PUBLISHER]

• Combination of deep learning network and clinical criteria ACR TI-RADS. • Large-scale and high-quality thyroid nodule ultrasound image dataset. • Efficient and lightweight network. Thyroid nodule is one of the most common endocrine diseases in the adult population. Early screening and diagnosis of thyroid nodules is of great significance to patients' subsequent treatment. An objective and accurate diagnosis algorithm for thyroid nodules is vital to improve the efficiency of clinical diagnosis and reduce the work pressure of doctors. In this paper, we propose a fused deep learning model for the diagnosis of benign and malignant thyroid nodules. Based on the ACR TI-RADS proposed by the American Academy of Radiology, the model first extracts 33 clinically significant statistical features of composition, echogenicity, shape, margin, and echogenic foci. After PCA dimensionality reduction, the top 4 features are fused with the feature map of EfficientNet B3. The thyroid nodules are classified into benign and malignant based on the fused feature map. This model is trained and tested on 3828 clinically collected ultrasound images of thyroid nodules. It achieves an accuracy of 96.4% and an AUC of 0.965. Compared with traditional machine learning algorithms and state-of-the-art deep learning networks, the fused model we proposed has improved classification performance and amount of parameters, and is also clinically interpretable. [ABSTRACT FROM AUTHOR]

Online class imbalance learning is a new learning problem that combines the challenges of both online learning and class imbalance learning. It deals with data streams having very skewed class distributions. This type of problems commonly exists in real-world applications, such as fault diagnosis of real-time control monitoring systems and intrusion detection in computer networks. In our earlier work, we defined class imbalance online, and proposed two learning algorithms OOB and UOB that build an ensemble model overcoming class imbalance in real time through resampling and time-decayed metrics. In this paper, we further improve the resampling strategy inside OOB and UOB, and look into their performance in both static and dynamic data streams. We give the first comprehensive analysis of class imbalance in data streams, in terms of data distributions, imbalance rates and changes in class imbalance status. We find that UOB is better at recognizing minority-class examples in static data streams, and OOB is more robust against dynamic changes in class imbalance status. The data distribution is a major factor affecting their performance. Based on the insight gained, we then propose two new ensemble methods that maintain both OOB and UOB with adaptive weights for final predictions, called WEOB1 and WEOB2. They are shown to possess the strength of OOB and UOB with good accuracy and robustness. [ABSTRACT FROM AUTHOR]

Pattern recognition using surface Electromyography (sEMG) applied on prosthesis control has attracted much attention in these years. In most of the existing methods, the sEMG signal during the firmly grasped period is used for grasp classification because good performance can be achieved due to its relatively stable signal. However, using the only the firmly grasped period may cause a delay to control the prosthetic hand gestures. Regarding this issue, we explored how grasp classification accuracy changes during the reaching and grasping process, and identified the period that can leverage the grasp classification accuracy and the earlier grasp detection. We found that the grasp classification accuracy increased along the hand gradually grasping the object till firmly grasped, and there is a sweet period before firmly grasped period, which could be suitable for early grasp classification with reduced delay. On top of this, we also explored corresponding training strategies for better grasp classification in real-time applications. [ABSTRACT FROM AUTHOR]

[Display omitted] • A low-cost fabric strain sensor with a spiral wound structure has been developed. • The sensor exhibits excellent linear sensing characteristics. • The relative resistance changes are consistent under different loading rates. • The sensor possesses broad sensing range, fast response, and high stability. This work presents a strain sensor which is based on high ductility liquid metal coated on traditional commercial lastex yarns with a unique spiral wound structure. The prepared sensor exhibits exceptional flexibility, elasticity, softness of 0.23 N at 100% tensile strain. It is lightweight, weighing only 1 g/m, and boasts a low cost of less than 1 dollar. The sensor has the ability to detect both static and dynamic in-plane tension and normal pressure. It showcases remarkable sensing characteristics of good sensitivity (0.82 for tension, 12.6 N−1 for pressure), extraordinary linearity within 10–100% strain range with goodness-of-fit value over 0.997, rapid response of less than 10 ms, high resolution spanning up to 0.3%, and high durability over 100,000 repetitive cycles. A smart glove based on the proposed strain sensors achieves gesture recognitions using machine learning technique, with an accuracy of 100%. The outstanding performance of this sensor highlights its potential as an affordable and advanced candidate for wearable electronics, which can have wide-ranging applications in sports, healthcare, entertainment, education, and industry. [ABSTRACT FROM AUTHOR]