Minimum Redundancy Maximum Relevance Algorithm Research Articles

Feasibility of a Clinical-Radiomics Model to Predict the Outcomes of Acute Ischemic Stroke.

ObjectiveTo develop a model incorporating radiomic features and clinical factors to accurately predict acute ischemic stroke (AIS) outcomes.Materials and MethodsData from 522 AIS patients (382 male [73.2%]; mean age ± standard deviation, 58.9 ± 11.5 years) were randomly divided into the training (n = 311) and validation cohorts (n = 211). According to the modified Rankin Scale (mRS) at 6 months after hospital discharge, prognosis was dichotomized into good (mRS ≤ 2) and poor (mRS > 2); 1310 radiomics features were extracted from diffusion-weighted imaging and apparent diffusion coefficient maps. The minimum redundancy maximum relevance algorithm and the least absolute shrinkage and selection operator logistic regression method were implemented to select the features and establish a radiomics model. Univariable and multivariable logistic regression analyses were performed to identify the clinical factors and construct a clinical model. Ultimately, a multivariable logistic regression analysis incorporating independent clinical factors and radiomics score was implemented to establish the final combined prediction model using a backward step-down selection procedure, and a clinical-radiomics nomogram was developed. The models were evaluated using calibration, receiver operating characteristic (ROC), and decision curve analyses.ResultsAge, sex, stroke history, diabetes, baseline mRS, baseline National Institutes of Health Stroke Scale score, and radiomics score were independent predictors of AIS outcomes. The area under the ROC curve of the clinical-radiomics model was 0.868 (95% confidence interval, 0.825–0.910) in the training cohort and 0.890 (0.844–0.936) in the validation cohort, which was significantly larger than that of the clinical or radiomics models. The clinical radiomics nomogram was well calibrated (p > 0.05). The decision curve analysis indicated its clinical usefulness.ConclusionThe clinical-radiomics model outperformed individual clinical or radiomics models and achieved satisfactory performance in predicting AIS outcomes.

A combined model for short-term wind power forecasting based on the analysis of numerical weather prediction data

Due to the fluctuation and intermittency of wind power resources, large-scale wind power integration brings serious challenges to power systems. Among the existing short-term forecasting methods, the accuracy and forecast period length can hardly meet the demand of power system economic dispatching and day-ahead power purchase markets. To further enhance the accuracy and increase the time scale, a short-term wind power forecasting (WPF) combined model based on numerical weather prediction (NWP) analysis is presented in this paper. First, according to the criterion of the minimum redundancy maximum relevance (mRMR) algorithm, several factors are sifted from the NWP multivariate data. Second, different characteristics of the factors are extracted, and weather patterns are divided into different types on the basis of these characteristics. Third, two deep learning models, Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), are applied for short-term WPF respectively under different weather types. Furthermore, the forecast results of the two models are combined using the Induced Ordered Weighted Average (IOWA) operator. The actual data collected from a wind farm in Northwest China are used to verify the conclusions. Results show that the proposed method can forecast wind power under different weather circumstances and outperform existing Radial Basis Function (RBF), Extreme Learning Machine (ELM) and Support Vector Machine (SVM)-based methods with respect to forecasting accuracy.

A retinal deep phenotypingTM platform to predict the cerebral amyloid PET status in older adults

AbstractBackgroundAs the only optically accessible part of the central nervous system, the retina represents an intriguing opportunity for the detection of biomarkers for Alzheimer’s disease (AD). This study evaluated the performance of the Retinal Deep PhenotypingTM platform, a digital biomarker platform comprising a hyperspectral retinal camera and image analysis algorithms, for the detection of likely positron‐emission tomography (PET) amyloid status (negative or positive) in older adults. A set of phenotypic features that correlates with the cerebral amyloid status as determined by amyloid PET scan were identified and used to train a classifying algorithm.MethodHyperspectral retinal images acquired with a Mydriatic Hyperspectral Retinal Camera from 194 participants (age ≥ 50 years), including cognitively normal and cognitively impaired (mild cognitive impairment and dementia) across 5 imaging sites were processed in order to train the model. Of these 194 participants, 73 individuals (38%) were amyloid‐positive, as confirmed by unanimous readings of PET scans by a panel of 3 expert reviewers. The pre‐processed hyperspectral images were segmented into various anatomical sites, and a texture‐based approach was used to extract several thousands of spatial‐spectral features. The most relevant features for the classification task were selected using a minimum redundancy maximum relevance (MRMR) algorithm and used to train a linear support vector machine (SVM) classifier. A nested, cross‐validation technique was used to evaluate the performance of the classifier.ResultThe resulting model based on the 17 most significant features showed high performance to discriminate between amyloid positive and negative subjects with an area under the receiver operating curve (AUCROC) of 0.87 (95% CI: 0.83 – 0.92).ConclusionThe Retinal Deep PhenotypingTM platform shows promise for detecting the likely cerebral amyloid PET status in adults 50 years and older from a simple, non‐invasive retinal scan and could provide an accessible means to identify individuals with abnormal cerebral amyloid in a clinical or drug development context. This phenotyping platform provides a flexible approach that could also be used for the detection of multiple biomarkers involved in cognitive decline from the same hyperspectral images of the retina.

Classification of COVID-19 by using supervised optimized machine learning technique

In recent two years, covid-19 diseases is the most harmful diseases in entire world. This disease increase the high mortality rate in several developed countries. Earlier identification of covid-19 symptoms can avoid the over illness or death. However, there are several researchers are introduced different methodology to identification of diseases symptoms. But, identification and classification of covid-19 diseases is the difficult task for every researchers and doctors. In this modern world, machine learning techniques is useful for several medical applications. This study is more focused in applying machine learning classifier model as SVM for classification of diseases. By improve the classification accuracy of the classifier by using hyper parameter optimization technique as modified cuckoo search algorithm. High dimensional data have unrelated, misleading features, which maximize the search space size subsequent in struggle to process data further thus not contributing to the learning practise, So we used a hybrid feature selection technique as mRMR (Minimum Redundancy Maximum Relevance) algorithm. The experiment is conducted by using UCI machine learning repository dataset. The classifier is conducted to classify the two set of classes such as COVID-19, and normal cases. The proposed model performance is analysed by using different parametric metrics, which are explained in result section.

Diagnosis of air compressor condition using minimum redundancy maximum relevance (MRMR) algorithm and distance metric based classification

Diagnosis of air compressor condition using minimum redundancy maximum relevance (MRMR) algorithm and distance metric based classification

MRI-Derived Radiomics of Hemophilic Arthropathy Is Associated with Severity of Joint Disease

IntroductionJoint bleeding resulting in synovial hypertrophy and articular cartilage damage are the hallmarks of hemophilic arthropathy. Early prophylaxis with extended half-life factor products and non-factor replacement therapies reduce joint bleeds but do not provide complete protection from development of arthropathy. MRI is the gold standard imaging technique to diagnose and monitor arthropathy in hemophilia. MRI changes that are associated with joint bleeding and development of arthropathy, however, are not sensitive enough to detect early changes in the synovial membranes after sub-clinical bleeding. Radiomics, the computerized extraction of sub-visual attributes from imaging scans can detect early changes of hemophilic arthropathy and help in more accurate characterizations of severe joint disease. The goal of this project was to develop and compare MRI radiomics with the International Prophylaxis Study Group (IPSG) score, an established MRI joint disease scale, in assessing the severity of joint disease.MethodsAfter IRB approval, knee MRI scans of 17 hemophilia patients with a history of joint bleeding and 12 age-matched healthy controls were included in this study. Manual annotations were performed in consensus by a board-certified musculoskeletal radiologist to include the prefemoral fat, suprapatellar fat pad and bursa . To ensure that the signal intensities from different cases were in tissue-specific correspondence, a landmark-based histogram transformation was used to align MRI signal intensity distributions across all cases. 560 radiomic features capturing MRI pixels texture heterogeneity (Gabor wavelets, Laws texture, Haralick and GoLIAGe co-occurrence patterns) were subsequently extracted on all MRI slices with ROI for all the cases. Top 2 radiomic features (F 2) for differentiating between abnormal joint changes and control were selected by Minimum Redundancy Maximum Relevance (mRMR) algorithm with 100 iterations of 3-fold cross-validation. Consensus hierarchical clustering was then applied on F 2, from which two patient clusters were generated. A disease severity score was calculated based on historical lifetime bleed events in the joint, designation of target joint status, and history of surgical synovectomy. A previously described 17-point MRI IPSG score was also calculated by 2 independent reviewers. Both the IPSG score and the F 2 were then fit into a linear regression model separately as predictors whereas the disease severity score was set to be the response variable. The Akaike's Information Criteria (AIC) and the root mean square error (RMSE) were used to evaluate the model fitness and the likelihood ratio (LR) test was applied for model comparison.ResultsThe dataset consisted of 17 males with moderate and severe hemophilia ranging in age from 10-55 years. The median disease severity score was 2 (0-11) and the median IPSG MRI Score was 2 (0-11). The top 2 radiomic features F 2, a Gabor wavelet feature and a Lawshc texture feature, yielded an AUC of 0.94 and an accuracy of 0.9 for differentiating between hemophilic joint changes and healthy control patients. Unsupervised clustering demonstrated good separation between the two classes (Fig panel A and B). The radiomic-based linear regression model (AIC: 86.67; RMSE: 2.86) was significantly better than the IPSG score-based model (AIC: 94.24; RMSE: 3.66) in joint disease severity characterizations (LR, p = 0.0083, Fig panel C).ConclusionsRadiomic analysis of knee MRI in hemophilia patients with history of joint bleeding is correlated with the severity of joint disease. Our study demonstrates the feasibility of developing and utilizing a radiomics based tool to detect the severity of joint damage in a small population of knee joint hemophilic arthropathy. We plan to perform an Independent validation of the radiomics signature in a larger dataset of hemophilia joint MRI scans [Display omitted] DisclosuresMadabhushi: Astrazeneca: Membership on an entity's Board of Directors or advisory committees; Bristol Meyers-Squibb: Membership on an entity's Board of Directors or advisory committees; Merck: Membership on an entity's Board of Directors or advisory committees; Roche: Consultancy; Caris: Consultancy; Aiforia: Consultancy; Philips: Research Funding; AstraZeneca: Research Funding; Boehringer-Ingelheim: Research Funding; Bristol Meyers-Squibb: Research Funding. Ahuja: Sanofi: Membership on an entity's Board of Directors or advisory committees; XaTek, Inc: Patents & Royalties; Genentech: Membership on an entity's Board of Directors or advisory committees; Takeda: Other: DSMB member .

Automated book location and classification method using RFID tags for smart libraries

PurposeThere are hundreds of thousands of books in libraries, causing problems in searching and counting books. To overcome these problems, this research presents a received signal strength indicator (RSSI) dataset and a new automated book location classification model using the RSSI dataset. Materials and methodIn this study, a RSSI signals dataset was collected using a mobile vehicle from 3279 books. These books are randomly placed on six benches, 24 cabinets, and 144 racks. The primary objective of this work is to detect book position automatically using a simple learning model. Thus, a new automated book position detection model is presented and this model is tested on the RSSI dataset collected. (i) Multiscale principal component analysis (MSPCA) is considered a pre-processing method. (ii) Iterative minimum redundancy maximum relevance (ImRMR) algorithm was used to select the most informative features automatically. (iii) The most informative features selected are classified using the Decision Trees (DT) and Ensemble Bagged Trees (BT) algorithms. Three cases are defined according to bench, cabinet, and rack. ResultsThe recommended model yielded 97.34% 98.26% and 93.1% overall accuracies using the bench, cabinet, and rack cases consecutively deploying the BT classifier with 10-fold cross-validation. ConclusionsThe presented book position classification model reached above 90% for all cases and these results calculated and finding clearly demonstrates that this model is ready to use in a big library for book position detection.

Unsupervised Minimum Redundancy Maximum Relevance Feature Selection for Predictive Maintenance

Identifying and selecting optimal prognostic health indicators in the context of predictive maintenance is essential to obtain a good model and make accurate predictions. Several metrics have been proposed in the past decade to quantify the relevance of those prognostic parameters. Other works have used the well-known minimum redundancy maximum relevance (mRMR) algorithm to select features that are both relevant and non-redundant. However, the relevance criterion is based on labelled machine malfunctions which are not always available in real life scenarios. In this paper, we develop a prognostic mRMR feature selection, an adaptation of the conventional mRMR algorithm, to a situation where class labels are a priori unknown, which we call unsupervised feature selection. In addition, this paper proposes new metrics for computing the relevance and compares different methods to estimate redundancy between features. We show that using unsupervised feature selection as well as adapting relevance metrics with the dynamic time warping algorithm help increase the effectiveness of the selection of health indicators for a rotating machine case study.

Radiomic features in recurrent/metastatic platinum refractory head and neck squamous cell carcinoma treated with immunotherapy.

e18000 Background: Radiomics is the computerized extraction of quantitative features from medical images, beyond the level of detail accessible to an unaided human eye. Several studies on radiomic analysis have been carried out to identify predictive, prognostic and diagnostic biomarkers for diverse tumor types including HNSCC. Radiomic features proved to be effective in predicting outcomes in patients with locally advanced HNSCC. The aim of the present study was to dentify predictive and prognostic radiomic features in platinum-refractory HNSCC patients with recurrent and/or metastatic (R/M) disease treated with anti PD-1 monotherapy. Methods: We retrospectively reviewed the data of 38 patients treated with Nivolumab at our Institution between January 2018 and March 2020. Nasopharyngeal carcinomas were not eligible. CT radiomic textural features were extracted from regions of interests (ROIs) manually delineated around tumor volumes. Minimum Redundancy Maximum Relevance (mRMR) algorithm was the method of choice to rank radiomic features based on three outcomes: overall response rate (ORR); disease progression (PD) and overall survival (OS). Logistic regression was employed to build predictive/prognostic models. Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) was chosen as the metric to assess model performances. Results: Data from 29 out of 38 patients were ultimately analyzed. Nine patients were excluded due to nasopharynx as primary tumor site and/or inadequate CT scan imaging. A total of 57 ROIs were extracted and analyzed. We obtained 9 models: 3 radiomic models, 3 clinical models and 3 combined models (radiomic plus clinical) for each outcome of interest (ORR, PD, OS). In the radiomic models 2 features were identified as predictor for ORR (AUC 0.69), 1 radiomic feature was predictive for PD (AUC 0.83) and 1 predictive for OS (AUC 0.72). In the clinical models, 3 clinical characteristics were identified for both ORR (AUC 0.91) and PD (AUC 0.73) and 2 clinical features were found to be predictive for OS (AUC 0.91). The combined model identified 3 features (2 radiomics and 1 clinical) predictive for ORR (AUC 0.76). No clinical characteristics were found in the combined model for PD. No radiomics features have been shown to be related to OS. Conclusions: This is an explorative study to test the power of radiomics and the potential value of combining radiomics and clinical data as a prognostic and predictive instrument. In this small sample size, the quantitative analysis of CT images seems to be an interesting tool which has to be further explored as predictor of outcome in a larger series of patients.[Table: see text]

Development of a Novel Web Camera-Based Contact-Free Major Depressive Disorder Screening System Using Autonomic Nervous Responses Induced by a Mental Task and Its Clinical Application.

BackgroundTo increase the consultation rate of potential major depressive disorder (MDD) patients, we developed a contact-type fingertip photoplethysmography-based MDD screening system. With the outbreak of SARS-CoV-2, we developed an alternative to contact-type fingertip photoplethysmography: a novel web camera-based contact-free MDD screening system (WCF-MSS) for non-contact measurement of autonomic transient responses induced by a mental task.MethodsThe WCF-MSS measures time-series interbeat intervals (IBI) by monitoring color tone changes in the facial region of interest induced by arterial pulsation using a web camera (1920 × 1080 pixels, 30 frames/s). Artifacts caused by body movements and head shakes are reduced. The WCF-MSS evaluates autonomic nervous activation from time-series IBI by calculating LF (0.04–0.15 Hz) components of heart rate variability (HRV) corresponding to sympathetic and parasympathetic nervous activity and HF (0.15–0.4 Hz) components equivalent to parasympathetic activities. The clinical test procedure comprises a pre-rest period (Pre-R; 140 s), mental task period (MT; 100 s), and post-rest period (Post-R; 120 s). The WCF-MSS uses logistic regression analysis to discriminate MDD patients from healthy volunteers via an optimal combination of four explanatory variables determined by a minimum redundancy maximum relevance algorithm: HF during MT (HFMT), the percentage change of LF from pre-rest to MT (%ΔLF(Pre–R⇒MT)), the percentage change of HF from pre-rest to MT (%ΔHF(Pre–R⇒MT)), and the percentage change of HF from MT to post-rest (%ΔHF(MT⇒Post–R)). To clinically test the WCF-MSS, 26 MDD patients (16 males and 10 females, 20–58 years) were recruited from BESLI Clinic in Tokyo, and 27 healthy volunteers (15 males and 12 females, 18–60 years) were recruited from Tokyo Metropolitan University and RICOH Company, Ltd. Electrocardiography was used to calculate HRV variables as references.ResultThe WCF-MSS achieved 73% sensitivity and 85% specificity on 5-fold cross-validation. IBI correlated significantly with IBI from reference electrocardiography (r = 0.97, p < 0.0001). Logit scores and subjective self-rating depression scale scores correlated significantly (r = 0.43, p < 0.05).ConclusionThe WCF-MSS seems a promising contact-free MDD screening apparatus. This method enables web camera built-in smartphones to be used as MDD screening systems.

Combining Multiparametric MRI Radiomics Signature With the Vesical Imaging-Reporting and Data System (VI-RADS) Score to Preoperatively Differentiate Muscle Invasion of Bladder Cancer.

BackgroundThe treatment and prognosis for muscle-invasive bladder cancer (MIBC) and non-muscle-invasive bladder cancer (NMIBC) are different. We aimed to construct a nomogram based on the multiparametric MRI (mpMRI) radiomics signature and the Vesical Imaging-Reporting and Data System (VI-RADS) score for the preoperative differentiation of MIBC from NMIBC.MethodThe retrospective study involved 185 pathologically confirmed bladder cancer (BCa) patients (training set: 129 patients, validation set: 56 patients) who received mpMRI before surgery between August 2014 to April 2020. A total of 2,436 radiomics features were quantitatively extracted from the largest lesion located on the axial T2WI and from dynamic contrast-enhancement images. The minimum redundancy maximum relevance (mRMR) algorithm was used for feature screening. The selected features were introduced to construct radiomics signatures using three classifiers, including least absolute shrinkage and selection operator (LASSO), support vector machines (SVM) and random forest (RF) in the training set. The differentiation performances of the three classifiers were evaluated using the area under the curve (AUC) and accuracy. Univariable and multivariable logistic regression were used to develop a nomogram based on the optimal radiomics signature and clinical characteristics. The performance of the radiomics signatures and the nomogram was assessed and validated in the validation set.ResultsCompared to the RF and SVM classifiers, the LASSO classifier had the best capacity for muscle invasive status differentiation in both the training (accuracy: 90.7%, AUC: 0.934) and validation sets (accuracy: 87.5%, AUC: 0.906). Incorporating the radiomics signature and VI-RADS score, the nomogram demonstrated better discrimination and calibration both in the training set (accuracy: 93.0%, AUC: 0.970) and validation set (accuracy: 89.3%, AUC: 0.943). Decision curve analysis showed the clinical usefulness of the nomogram.ConclusionsThe mpMRI radiomics signature may be useful for the preoperative differentiation of muscle-invasive status in BCa. The proposed nomogram integrating the radiomics signature with the VI-RADS score may further increase the differentiation power and improve clinical decision making.

M-Cuckoo and SVM Classification Algorithm Based Opinion Mining

Opinion Mining or Sentiment Analysis is a task in the processing of natural language to find the customers' mood about buying a specific product or subject. It involves developing a framework in many online shopping sites to gather and review opinions about the product made. Opinion mining is a sub-field of the mining of web content. Data mining is a branch of Web content mining. Opinions are statements that reflect the opinion or sentiment of individuals. Opinion on objects or events is also given in this statement. For any person, reviewing consumer review is more relevant in making the right buying product and organization decision. CS is the best search algorithm inspired by cuckoos' breeding behavior. It provides a short overview of the nature-inspired algorithm's applications. The CS algorithm is used in various fields, such as business, image processing, wireless sensor networks, flood forecasting, document clustering, speaker recognition, distributed system shortest path, health sector, job scheduling. In terms of better efficiency and less processing time, the Cuckoo algorithm performs various nature-inspired algorithms. Therefore, this research paper proposes a hybrid feature selection which is a combination of cuckoo search and mRMR (Minimum Redundancy Maximum Relevance) algorithm. Due to the subjective nature of social media reviews, hybrid feature selection technique outperforms the traditional technique. The performance factors like f-measure, recall, precision, and accuracy tested on Amazon dataset using Support Vector Machine (SVM) classifier.

CT Texture Analysis for Preoperative Identification of Lymphoma from Other Types of Primary Small Bowel Malignancies.

Objectives To explore the application of computed tomography (CT) texture analysis in differentiating lymphomas from other malignancies of the small bowel. Methods Arterial and venous CT images of 87 patients with small bowel malignancies were retrospectively analyzed. The subjective radiological features were evaluated by the two radiologists with a consensus agreement. The region of interest (ROI) was manually delineated along the edge of the lesion on the largest slice, and a total of 402 quantified features were extracted automatically from AK software. The inter- and intrareader reproducibility was evaluated to select highly reproductive features. The univariate analysis and minimum redundancy maximum relevance (mRMR) algorithm were applied to select the feature subsets with high correlation and low redundancy. The multivariate logistic regression analysis based on texture features and radiological features was employed to construct predictive models for identification of small bowel lymphoma. The diagnostic performance of multivariate models was evaluated using receiver operating characteristic (ROC) curve analysis. Results The clinical data (age, melena, and abdominal pain) and radiological features (location, shape, margin, dilated lumen, intussusception, enhancement level, adjacent peritoneum, and locoregional lymph node) differed significantly between the nonlymphoma group and lymphoma group (p < 0.05). The areas under the ROC curve of the clinical model, arterial texture model, and venous texture model were 0.93, 0.92, and 0.87, respectively. Conclusion The arterial texture model showed a great diagnostic value and fitted performance in preoperatively discriminating lymphoma from nonlymphoma of the small bowel.

Brain MRI radiomics analysis may predict poor psychomotor outcome in preterm neonates.

This study aimed to apply a radiomics approach to predict poor psychomotor development in preterm neonates using brain MRI. Prospectively enrolled preterm neonates underwent brain MRI near or at term-equivalent age and neurodevelopment was assessed at a corrected age of 12 months. Two radiologists visually assessed the degree of white matter injury. The radiomics analysis on white matter was performed using T1-weighted images (T1WI) and T2-weighted images (T2WI). A total of 1906 features were extracted from the images and the minimum redundancy maximum relevance algorithm was used to select features. A prediction model for the binary classification of the psychomotor developmental index was developed and eightfold cross-validation was performed. The diagnostic performance of the model was evaluated using the AUC with and without including significant clinical and DTI parameters. A total of 46 preterm neonates (median gestational age, 29 weeks; 26 males) underwent brain MRI (median corrected gestational age, 37 weeks). Thirteen of 46 (28.3%) neonates showed poor psychomotor outcomes. There was one neonate among 46 with moderate to severe white matter injury on visual assessment. For the radiomics analysis, twenty features were selected for each analysis. The AUCs of prediction models based on T1WI, T2WI, and both T1WI and T2WI were 0.925, 0.834, and 0.902. Including gestational age or DTI parameters did not improve the prediction performance of T1WI. A radiomics analysis of white matter using early T1WI or T2WI could predict poor psychomotor outcomes in preterm neonates. • Radiomics analysis on T1-weighted images of preterm neonates showed the highest diagnostic performance (AUC, 0.925) for predicting poor psychomotor outcomes. • In spite of 45 of 46 neonates having no significant white matter injury on visual assessment, the radiomics analysis of early brain MRI showed good diagnostic performance (sensitivity, 84.6%; specificity, 78.8%) for predicting poor psychomotor outcomes. • Radiomics analysis on early brain MRI can help to predict poor neurodevelopmental outcomes in preterm neonates.

Integrating No.3 lymph nodes and primary tumor radiomics to predict lymph node metastasis in T1-2 gastric cancer

BackgroundThis study aimed to develope and validate a radiomics nomogram by integrating the quantitative radiomics characteristics of No.3 lymph nodes (LNs) and primary tumors to better predict preoperative lymph node metastasis (LNM) in T1-2 gastric cancer (GC) patients.MethodsA total of 159 T1-2 GC patients who had undergone surgery with lymphadenectomy between March 2012 and November 2017 were retrospectively collected and divided into a training cohort (n = 80) and a testing cohort (n = 79). Radiomic features were extracted from both tumor region and No. 3 station LNs based on computed tomography (CT) images per patient. Then, key features were selected using minimum redundancy maximum relevance algorithm and fed into two radiomic signatures, respectively. Meanwhile, the predictive performance of clinical risk factors was studied. Finally, a nomogram was built by merging radiomic signatures and clinical risk factors and evaluated by the area under the receiver operator characteristic curve (AUC) as well as decision curve.ResultsTwo radiomic signatures, reflecting phenotypes of the tumor and LNs respectively, were significantly associated with LN metastasis. A nomogram incorporating two radiomic signatures and CT-reported LN metastasis status showed good discrimination of LN metastasis in both the training cohort (AUC 0.915; 95% confidence interval [CI] 0.832–0.998) and testing cohort (AUC 0.908; 95% CI 0.814–1.000). The decision curve also indicated its potential clinical usefulness.ConclusionsThe nomogram received favorable predictive accuracy in predicting No.3 LNM in T1-2 GC, and the nomogram showed positive role in predicting LNM in No.4 LNs. The nomogram may be used to predict LNM in T1-2 GC and could assist the choice of therapy.

BREAST CANCER DETECTION USING RSFS-BASED FEATURE SELECTION ALGORITHMS IN THERMAL IMAGES

Breast cancer is a common cancer in female. Accurate and early detection of breast cancer can play a vital role in treatment. This paper presents and evaluates a thermogram based Computer-Aided Detection (CAD) system for the detection of breast cancer. In this CAD system, the Random Subset Feature Selection (RSFS) algorithm and hybrid of minimum Redundancy Maximum Relevance (mRMR) algorithm and Genetic Algorithm (GA) with RSFS algorithm are utilized for feature selection. In addition, the Support Vector Machine (SVM) and k-Nearest Neighbors (kNN) algorithms are utilized as classifier algorithm. The proposed CAD system is verified using MATLAB 2017 and a dataset that is composed of breast images from 78 patients. The implementation results demonstrate that using RSFS algorithm for feature selection and kNN and SVM algorithms as classifier have accuracy of 85.36% and 75%, and sensitivity of 94.11% and 79.31%, respectively. In addition, using hybrid GA and RSFS algorithm for feature selection and kNN and SVM algorithms as classifier have accuracy of 83.87% and 69.56%, and sensitivity of 96% and 81.81%, respectively, and using hybrid mRMR and RSFS algorithms for feature selection and kNN and SVM algorithms as classifier have accuracy of 77.41% and 73.07%, and sensitivity of 98% and 72.72%, respectively.

Propose new implement models to determine the compressive, tensile and flexural strengths of recycled coarse aggregate concrete via imperialist competitive algorithm

This study aims to develop new formulas to predict the tensile, compressive and flexural strengths of recycled coarse aggregates (RCA) concrete. For this purpose, a total of 1348 available experimental results are used. Imperialist Competitive Algorithm (ICA) is employed to find equations based on the content of water, cement, RCA, natural coarse aggregates (NCA) and natural fine aggregates (NFA). In addition, some relationships are developed between flexural, tensile and compressive strengths when different RCA contents are used. To provide the proper equations for estimating the compression, bending and tension capabilities, first the features are considered using the minimum redundancy maximum relevance algorithm that affects the efficiency of the algorithms. Then, the appropriate number of features for estimating each parameter is selected using the Multi-Layer Perceptron (MLP) network. The obtained results showed that the mean absolute error of the proposed formulas in estimating the compressive, flexural and tensile strengths are about 0.54, 0.36 and 0.48, respectively. Additionally, RCA had a substantial influence on the mechanical properties of concrete, and the RCA content should be considered in the formulas utilized to foresee the mechanical characteristics of concrete.

Efficient deep feature extraction and classification for identifying defective photovoltaic module cells in Electroluminescence images

Electroluminescence (EL) imaging has become the standard test procedure for defect detection throughout the production, installation and operation stages of solar modules. Using this test, defects such as micro cracks, broken cells, and finger interruptions on photovoltaic modules could be easily detected and potential power loss issues could be effectively addressed. Although EL test is a very powerful inspection method, interpreting the EL images could be quite challenging due to the inhomogeneous background and complex defect patterns. Therefore, evaluating the damaged cells and determining the defect severity require expertise, and could be time consuming to apply these processes manually for each cell. Hence, the automatic visual inspection of photovoltaic cells is very important. In this study, a novel automatic defect detection and classification framework for solar cells is proposed. In the proposed Deep Feature-Based (DFB) method, the image features extracted through deep neural networks are classified with machine learning methods such as support vector machines, K-Nearest Neighborhood, Decision Tree, Random Forest and Naive Bayes. Thus, classical machine learning and deep learning techniques are used together. In order to combine the features taken from different deep network architectures in various combinations, the minimum Redundancy Maximum Relevance (mRMR) algorithm is employed for the feature selection. In this way, the dimensions of the feature vectors are reduced and the classification performance is increased with fewer features. With the determination of the best features extracted from different layers of deep neural networks, state-of-the-art results were obtained for both 4-class and 2-class datasets. Moreover, a Lightweight Convolutional Neural Network (L-CNN) architecture has been proposed and trained from scratch, and the results are compared with previous works. As a result, the highest scores are obtained using DFB method with Support Vector Machines (SVM) and classification scores of 90.57% and 94.52% were obtained for the dataset with 4 - class and 2 - class, respectively. The proposed DFB-SVM models outperformed other studies using the same dataset. The results showed that the proposed framework can detect PV cell defects with high accuracy.

Computed Tomography-Based Radiomics Model for Predicting the WHO/ISUP Grade of Clear Cell Renal Cell Carcinoma Preoperatively: A Multicenter Study.

PurposeTo examine the ability of computed tomography radiomic features in multivariate analysis and construct radiomic model for identification of the the WHO/ISUP pathological grade of clear cell renal cell carcinoma (ccRCC).MethodsThis was a retrospective study using data of four hospitals from January 2018 to August 2019. There were 197 patients with a definitive diagnosis of ccRCC by post-surgery pathology or biopsy. These subjects were divided into the training set (n = 122) and the independent external validation set (n = 75). Two phases of Enhanced CT images (corticomedullary phase, nephrographic phase) of ccRCC were used for whole tumor Volume of interest (VOI) plots. The IBEX radiomic software package in Matlab was used to extract the radiomic features of whole tumor VOI images. Next, the Mann–Whitney U test and minimum redundancy-maximum relevance algorithm(mRMR) was used for feature dimensionality reduction. Next, logistic regression combined with Akaike information criterion was used to select the best prediction model. The performance of the prediction model was assessed in the independent external validation cohorts. Receiver Operating Characteristic curve (ROC) was used to evaluate the discrimination of ccRCC in the training and independent external validation sets.ResultsThe logistic regression prediction model constructed with seven radiomic features showed the best performance in identification for WHO/ISUP pathological grades. The Area Under Curve (AUC) of the training set was 0.89, the sensitivity comes to 0.85 and specificity was 0.84. In the independent external validation set, the AUC of the prediction model was 0.81, the sensitivity comes to 0.58, and specificity was 0.95.ConclusionA radiological model constructed from CT radiomic features can effectively predict the WHO/ISUP pathological grade of CCRCC tumors and has a certain clinical generalization ability, which provides an effective value for patient prognosis and treatment.

Ultrasound-based deep learning radiomics in the assessment of pathological complete response to neoadjuvant chemotherapy in locally advanced breast cancer

PurposeThe aim of the study was to develop and validate a deep learning radiomic nomogram (DLRN) for preoperatively assessing breast cancer pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) based on the pre- and post-treatment ultrasound. MethodsPatients with locally advanced breast cancer (LABC) proved by biopsy who proceeded to undergo preoperative NAC were enrolled from hospital #1 (training cohort, 356 cases) and hospital #2 (independent external validation cohort, 236 cases). Deep learning and handcrafted radiomic features reflecting the phenotypes of the pre-treatment (radiomic signature [RS] 1) and post-treatment tumour (RS2) were extracted. The minimum redundancy maximum relevance algorithm and the least absolute shrinkage and selection operator regression were used for feature selection and RS construction. A DLRN was then developed based on the RSs and independent clinicopathological risk factors. The performance of the model was assessed with regard to calibration, discrimination and clinical usefulness. ResultsThe DLRN predicted the pCR status with accuracy, yielded an area under the receiver operator characteristic curve of 0.94 (95% confidence interval, 0.91–0.97) in the validation cohort, with good calibration. The DLRN outperformed the clinical model and single RS within both cohorts (P < 0.05, as per the DeLong test) and performed better than two experts' prediction of pCR (both P < 0.01 for comparison of total accuracy). Besides, prediction within the hormone receptor–positive/human epidermal growth factor receptor 2 (HER2)–negative, HER2+ and triple-negative subgroups also achieved good discrimination performance, with an AUC of 0.90, 0.95 and 0.93, respectively, in the external validation cohort. Decision curve analysis confirmed that the model was clinically useful. ConclusionA deep learning–based radiomic nomogram had good predictive value for pCR in LABC, which could provide valuable information for individual treatment.