Computer Vision Approach for Liver Tumor Classification Using CT Dataset

Clinical decision making and research in hepatocellular carcinoma: Pivotal role of imaging techniques

Detection of visual faults in photovoltaic modules using a stacking ensemble approach

Recent advances in color Doppler have made it possible to demonstrate the blood flow signals within liver tumors [1, 2]. Some characteristic patterns in color Doppler have been demonstrated in malignant liver tumors, and characteristic findings in tumor invasion into the portal vein have been reported [3, 4]. However, the poor detectability of color Doppler signals prevents the wide use of this technique in the clinical investigation of liver tumors. Contrast enhancement is expected to improve the diagnostic ability of color Doppler in tumors. Carbon dioxide microbubbles have been already used for the B-mode enhancement of liver tumors, and allows precise evaluation of tumor vascularity [5]. However, it is available only by intra-arterial injection during hepatic angiography. Recently, new contrast agents which enhance the liver after intravenous injection have been developed. The enhancement of color Doppler signals in liver tumors after intravenous injection of sonicated albumin was observed in an experimental in vivo study [6]. In this article, recent clinical experience with a galactose-based contrast agent in the diagnosis of hepatocellular carcinoma (HCC) is presented and its significance for clinical practice is discussed.

To combine ultrasound-guided fine-needle aspiration (US-FNA) and Liu (Riu) stain to make a quick study on liver tumor lesions. Two hundred and twenty-eight aspirations from 232 patients were completely studied. The operator himself made the quick cytodiagnosis of US-FNA smear stained by Liu method within thirty minutes. The US-FNA specimen was also sent to the pathological department for cytological study and cellblock histology. The result of our Liu-stain quick cytodiagnosis in each patient was confirmed by the final cytopathological diagnosis from pathological report. Among 228 samples, the quick cytodiagnosis revealed 146 malignancies, 81 benign lesions and one inadequate specimen. Cytopathological diagnosis from the pathological department revealed 150 malignancies, and 78 benign lesions. Four well-differentiated hepatocellular carcinomas (HCCs) were under-diagnosed by quick cytodiagnosis as benign and 3 benign lesions were over-diagnosed as well-differentiated HCCs. Compared with cytopathological diagnosis, quick cytodiagnosis correctly diagnosed 143 malignancies and 77 benign lesions. Except for the one inadequate specimen in quick cytodiagnosis, the accuracy of quick cytodiagnosis was 96.9% (220/227), and its sensitivity, specificity and positive and negative predictive values were 97.9%, 95.1%, 97.3% and 96.3%, respectively. Liu-stain quick cytodiagnosis is a fast, convenient, safe and effective method for hepatologists in clinic practice to diagnose liver tumor. In few cases of well-differentiated HCC, Liu-stain quick cytodiagnosis has its limitation.

Diagnosing liver tumors can be difficult in the setting of a poorly differentiated tumor or tumors with no known prior malignancy. Frequently, alpha-fetoprotein, carcinoembryonic antigen, factor VIII, and mucicarmine have been employed to distinguish hepatocellular carcinoma (HCC) from adenocarcinoma. However, these stains have their limitations. CA 15-3 and Glut-1 are expressed in a variety of carcinomas. To the authors' knowledge, their expression in HCC has not been studied extensively. The authors examined the clinical utility of CA 15-3 and Glut-1 in the setting of fine-needle aspiration biopsy samples from the liver. Thirty-five cases of HCC and 59 cases of tumors metastatic to the liver were studied. These cases previously were studied with the hepatocyte paraffin-1 antibody. Each case was stained with CA 15-3 and Glut-1 using the avidin-biotin complex method. Each case was evaluated in a blinded fashion for membranous staining that was stronger than cytoplasmic or background staining. The diagnoses were unblinded and staining patterns were compared. CA 15-3 stained 43 of 59 metastatic carcinoma samples and 3 of 35 HCC samples. Glut-1 stained 34 of 59 metastases and 2 of 35 HCCs. Together, the 2 immunostains stained 51 of 59 metastases and 5 of 35 HCCs. Diagnostic accuracy was improved by adding hepatocyte paraffin-1 to the staining panel. CA 15-3 and Glut-1, especially in conjunction with hepatocyte paraffin-1, appear to be helpful in discriminating HCC from other carcinomas.

Objective To compare the diagnostic performances of real time of shear wave elastography (SWE) and strain ratio (SR) in differentiating benign and malignant liver tumors. Methods From March 2017 to August 2017, 68 patients with 70 lesions were examined with SR and SWE respectively, the parameters of SWE were maximum real-time shear wave velocity (Vmax), average real-time shear wave velocity (Vmean) and real-time shear wave velocity elasticity ratio of lesion/peripheral liver tissue (Vratio). The receiver operating characteristic (ROC) curve of SWE and SR parameters were plotted to assess the value in distinguishing benign and malignant liver tumors. The results were compared using the area under the ROC curve (AUC). The sensitivity, specificity, accuracy of SWE and SR were calculated, The accuracy of the parameters with the largest AUC were compared between SWE and SR by McNemar test in differentiating focal liver lesions. Results The AUC of SR, Vmax, Vmean and Vratio were 0.948, 0.868, 0.918 and 0.956 respectively. The comparison of AUC reveals no significant difference between them. The sensitivity, specificity, accuracy of SR in differentiating benign and malignant liver tumors were 92.73%, 93.33%, 92.86% respectively (P>0.05). The SWE parameter with the largest AUC was Vratio. The sensitivity, specificity, accuracy of Vratio was 94.56%, 100.00%, 95.71%, respectively. The accuracy of Vratio and SR in differentiating benign and malignant liver tumors showed no significant difference (P>0.05). Conclusions The application of SWE and SR in the differential diagnosis of benign and malignant liver tumors is of high diagnostic value. They provide a new method for the differential diagnosis of liver tumor lesions, and the diagnostic value of both of them is equivalent. Key words: Liver neoplasms/US; Elasticity imaging techniques; Diagnosis, differential

Entrepreneur and entrepreneurship are two words profoundly involved in every sector of the world. In this era of modern economy, the most important thing is that entrepreneurs are the main driving force of the global economy. Several kinds of entrepreneurship ventures are emerging with the help of technological support. With the help of information technology and the access to the new generation customers, the ventures are outperforming the old-fashioned companies. The startup companies nowadays are started by young students and professionals. In this paper, we have focused on tertiary-level students, who are currently at their studies and thinking about starting a startup company. We have collected self-report inventories for determining personality traits and entrepreneurship skill level. The Big Five personality model and entrepreneurship self-assessment survey for collecting data. With the machine learning models, we have tried to correlate personality traits and entrepreneurship skills, therefore, to find out the actual skills of entrepreneurs. Around one hundred participants have participated in this research. For the machine learning classification, we have applied tree-based algorithms such as, decision tree (J48), random forest (RF), decision stump, Hoeffding tree, random tree, logistic model tree (LMT) and REP tree. The performance evaluation of the classifiers is performed considering three classes, namely outstanding ability to be an entrepreneur, satisfactory ability to be an entrepreneur and inappropriate for entrepreneurship. Among the classifiers applied, Hoeffding tree performed better (88%), in terms of accuracy.KeywordsBig Five personality traitsEntrepreneurial skillMachine learning techniques

The area of corporate bankruptcy prediction attains high economic importance, as it affects many stakeholders. The prediction of corporate bankruptcy has been extensively studied in economics, accounting and decision sciences over the past two decades. The corporate bankruptcy prediction has been a matter of talk among academic literature and professional researchers throughout the world. Different traditional approaches were suggested based on hypothesis testing and statistical modeling. Therefore, the primary purpose of the research is to come up with a model that can estimate the probability of corporate bankruptcy by evaluating its occurrence of failure using different machine learning models. As the dataset was not well prepared and contains missing values, various data mining and data pre-processing techniques were utilized for data preparation. Within this research, the task of resolving the issues induced by the imbalance between the two classes is approached by applying different data balancing techniques. We address the problem of imbalanced data with the random undersampling and Synthetic Minority Over Sampling Technique (SMOTE). We used five machine learning models (support vector machine, J48 decision tree, Logistic model tree, random forest and decision forest) to predict corporate bankruptcy earlier to the occurrence. We use data from 2009 to 2013 on Poland manufacturing corporates and selected the 64 financial indicators to be broken down. The main finding of the study is a significant improvement in predictive accuracy using machine learning techniques. We also include other economic indicators ratios, along with Altman’s Z-score variables related to profitability, liquidity, leverage and solvency (short/long term) to propose an efficient model. Machine learning models give better results while balancing the data through SMOTE as compared to random undersampling. The machine learning technique related to decision forest led to 99% accuracy, whereas support vector machine (SVM), J48 decision tree, Logistic Model Tree (LMT) and Random Forest (RF) led to 92%, 92.3%, 93.8% and 98.7% accuracy, respectively, with all predictive financial indicators. We find that the decision forest outperforms the other techniques and previous techniques discussed in the literature. The proposed method is also deployed on the web to assist regulators, investors, creditors and scholars to predict corporate bankruptcy.

Analyzing the hidden information from the images are helpful to identify the various causes. In general processing of images includes Pre-processing, segmentation, Feature extraction and Classification. Significance of classifier is essential since results are always based on the classifier. The ultimate aim is to explain how WEKA tool is used for rheumatoid arthritis and investigate the performance of the various classifiers for huge data. In our method we are distinctively give attention to the classification methods like ADTree, Best First Decision tree(BF), Decision stump, J48Pruned tree, J48 Graft Pruned tree, Least Absolute Deviation regression trees (LAD), Logistic Model Tree(LMT), Naïve-Bayes (NB), Random tree, Random forest tree, CART Decision tree. The features like Area, perimeter, circularity, integrated density, Median, Skewness, Raw integrated density, and Roundness and solidity are obtained from the Lymphocytes images and formed the data set. Different classifier is applied for RA facet for Validation. RA facet contains 108 rows and 10 columns. Using classifier to find out the various values like Relative Absolute Error, and Kappa Statistic, Mean Absolute Error, Root Mean Squared Error and Root Relative Squared Error. From those values compare with all the methods ADT classifier is suggested for use in huge data.

Accurate classification of Turkish voice commands is essential for advancing voice-controlled technologies and enabling seamless human-computer interaction in native language contexts. This study systematically evaluates multiple feature extraction models capturing temporal, spectral, and time-frequency characteristics of speech signals to enhance recognition accuracy. Six feature vector models were developed, with the final model integrating Information Gain-based feature selection and Linear Predictive Coding-derived formant frequencies to create a comprehensive and discriminative representation. Classification was performed using six widely adopted algorithms: Random Forest, k-Nearest Neighbors, Multilayer Perceptron, Logistic Model Tree, Support Vector Machine, and an Ensemble voting method combining Random Forest, Multilayer Perceptron, and Logistic Model Tree. The Ensemble voting classifier demonstrated superior performance, achieving an accuracy of 93.94%, significantly outperforming individual classifiers and baseline models. This study contributes to the literature by presenting a robust, explainable, and high-performing feature framework tailored for Turkish voice command recognition. The integration of spectral, temporal, and articulatory features enables improved discrimination of speech commands, offering valuable insights for future voice-activated applications in Turkish language environments.

A new hybrid classification algorithm for customer churn prediction based on logistic regression and decision trees

Based on data throughout 2022, there have been many reductions in employees both globally and Indonesia. The reduction was made due to adjustments with developments to keep the business afloat in increasingly fierce competition. However, reducing the number of employees is not an easy decision to make. This decision can have an impact on many aspects of the development and course of a business or company. To make a decision especially related to the aspect of termination of employment, it is necessary to consider carefully and thoroughly. Assessment and decision-making cannot be based on just one aspect, other aspects need to be seen to be taken into consideration. Additional aspects that can be selected to strengthen decision-making can be taken from the data. Data will not have any value without processing it with various approaches, one of which is the prediction process. Starting from the data, the prediction results will be more appropriate to make a decision. This study made a comparison of 3 decision tree algorithms, and produced a comparison of the three methods in terms of accuracy. The results of this study are the best accuracy for each algorithm C.45 = 83.44; Random Forests = 85.85; LMT = 88.29 with a linear precision value, and the best algorithm model with the highest accuracy is the Logistic Model Tree (LMT) algorithm.

Classification of team sport activities using a single wearable tracking device

X-ray computed tomography (CT) imaging plays an essential role in disease diagnosis due to its noninvasive, painless mode and superior penetration depth. However, the resolution of the soft tissue ...

One of the major shares of the current security infrastructure is formed by the Intrusion Detection Systems (IDS). The attack launched towards the security systems are increasing in a rapid way. The sophistication of attack methods with more automated tools enables the attackers to gain control over the systems and produce threats to the information assets. The normal way of detecting the attacks is by using tools that produce alerts to the system administrators. But most of the attacks would normally escape from these tools since they are mostly rule-based. So the need for enhanced attack detection methods becomes vital for the security infrastructure. The attack detection methods are normally statistical based or probabilistic based. This paper focuses on attack detection using multiple learning based classifiers such as J48, Naïve Bayes, Random Forest, Random Tree, KStar, RotationForest, RandomSubspace, Ordinal Class Classifier, Data Near BalancedND and Multiclass classifier. Correlation Based Feature Selection (CFS) is also used to select the best features of the kddcup 99 dataset for the attack classes such as DoS, Probe, U2R and R2L. The feature selection enables the classifiers to improve the accuracy of classification. The multiple classifiers are used in four layers for detecting the four types of attack classes. The classification rate of above 99% is obtained. Cost - Benefit analysis is done for the various attack detection methods and the ROC curves are also plotted.