The problem of interactive foreground/background segmentation in still images is of great practical importance in image editing. They avoid the boundary-length bias of graph-cut methods and results in increased sensitivity to seed placement. A new proposed method of fully automatic processing frameworks is given based on Graph-cut and Geodesic Graph cut algorithms. This paper addresses the problem of segmenting liver and tumor regions from the abdominal CT images. The lack of edge modelling in geodesic or similar approaches limits their ability to precisely localize object boundaries, something at which graph-cut methods generally excel. A predicate is defined for measuring the evidence for a boundary between two regions using Geodesic Graph-based representation of the image. The algorithm is applied to image segmentation using two different kinds of local neighborhoods in constructing the graph. Liver and hepatic tumor segmentation can be automatically processed by the Geodesic graph-cut based method. This system has concentrated on finding a fast and interactive segmentation method for liver and tumor segmentation. In the pre-processing stage, Mean shift filter is applied to CT image process and statistical thresholding method is applied for reducing processing area with improving detections rate. In the Second stage, the liver region has been segmented using the algorithm of the proposed method. Next, the tumor region has been segmented using Geodesic Graph cut method. Results show that the proposed method is less prone to shortcutting than typical graph cut methods while being less sensitive to seed placement and better at edge localization than geodesic methods. This leads to increased segmentation accuracy and reduced effort on the part of the user. Finally Segmented Liver and Tumor Regions were shown from the abdominal Computed Tomographic image.

Heart disease is the number one cause of death for all communities of individuals in advanced countries and a major problem for emerging nations too. Doctors' availability to care for the general population could not catch up with the present demand for healthcare. So, there is a severe need for a support system to assist save individuals. With novel ML frameworks and big data repositories, our motive is to design a machine learning model to predict heart disease at the earliest, help prioritize hospital consultations and improve accuracy. For this study, several analyzes were carried out on the Cleveland heart disease data set with 303 patients records, using five different classifiers namely Support Vector Machine (SVM), Random forests, Ordinal Regression, Logistic Regression and Naïve Bayes. Feature selection using chi-squared statistical test and correct tuning of hyperparameters maximized classification accuracy of the Support vector machine (Radial basis function) from 40% to 85%. By incorporating rules based on the statistical patterns observed, the efficiency was further enhanced to 95%. On the other side, seeing it as a 5-class classification, multi-class imbalance issue was addressed using suitable sampling techniques that resulted in 96% accuracy for 5-class data. We evaluated model efficiency using k-fold cross validation and confusion matrix. This study shows that the classification accuracy could be significantly improved by balancing the dataset using sampling and by properly tuning hyperparameters after feature selection.

No abstract

Datasets may have large number of features which makes it hard and time consuming to classify. Additionally, they may have irrelevant and noise features too with missing values. The missing values should be treated in a proper way so that the classifier accuracy can be improved. There is also a need to reduce features and select only the features necessary to the classifier. Principal Component Analysis (PCA) is commonly considered for this process of reducing the number of features in a dataset. These reduced components can be applied as input to the classifiers. In this study, standard datasets are checked for missing values, classified using Support vector Machines (SVM) and Naive Bayes with and without reducing the features using PCA. Then, the proposed algorithm for missing value imputation is used on the datasets and the same analysis were carried out. The accuracy is evaluated using Confusion Matrix. The results are discussed with analysis based on the nature of features and missing values and how different datasets behave when used with machine learning algorithms.

Visual Speech Recognition aims at transcribing lip movements into readable text. There have been many strides in automatic speech recognition systems that can recognize words with audio and visual speech features, even under noisy conditions. This paper focuses only on the visual features, while a robust system uses visual features to support acoustic features. We propose the concatenation of visemes (lip movements) for text classification rather than a classic individual viseme mapping. The result shows that this approach achieves a significant improvement over the state-of-the-art models. The system has two modules; the first one extracts lip features from the input video, while the next is a neural network system trained to process the viseme sequence and classify it as text.