Search

An abundance of various segmentation techniques are available in the literature, that cater to wide range of image understanding applications. The paper proposes a unified way of systematic categorization of the research work on image segmentation called Multi-Faceted Hierarchical Image Segmentation Taxonomy (MFHIST), which consist of six facets presented in a hierarchical manner - scope, requirement, control, feature, image representation and approach specifications. Every scope is exemplified with research works from the literature for better understanding. The paper gives an illustration of populating MFHIST, to provide the reader a quick grasp of few important state-of-art image segmentation research works and their adaptations. As a case study, the illustrations display a limited version to uncover the journey of basic to modern adaptations in the areas region based segmentation approach, such as Markov Random Fields, Spectral Clustering, Active Contour Model, Mean Shift Clustering. The other segmentation approaches have not been considered here, owing to the enormous volume of works in the past four to five decades and limitation in articulating all of them using MFHIST. The performance analysis of the algorithms using quantitative metrics is not in the present scope and will be considered in future version of MFHIST.

Machine learning (ML) is accurate and reliable in solving supervised problems such as classification, when the training is performed appropriately for the predefined classes. In real world scenario, during the dataset creation, class imbalance may arise, where one of the classes has huge number of instances while the other class has very less in numbers. In other words, the class distribution is not equal. Such scenarios results in anomalous prediction result. Handling of imbalanced dataset is therefore required to make correct prediction considering all the class scenarios in an equal ratio. The paper mentions various external and internal techniques to balance dataset found in literature survey along with experimental analysis of four different datasets from various domains- medical, mining, security, finance. The experiments are done using Python. External balancing techniques are used to balance the datasets- two oversampling SMOTE and ADASYN techniques and two undersampling Random Undersampling and Near Miss techniques. These datasets are used for binary classification task. Three machine learning classification algorithms such as logistic regression, random forest and decision tree are applied to imbalanced and balanced datasets to compare and contrast the performances. Comparisons with both balanced and unbalanced are reported. It has been found that undersample technique loses many important datapoints and thereby predicts with low accuracy. For all the datasets it is observed that oversampling technique ADASYN makes some decent prediction with appropriate balance.

This paper proposes a model based on convolutional neural network for hand gesture recognition and classification. The dataset uses 26 different hand gestures, which map to English alphabets A–Z. Standard dataset called Hand Gesture Recognition available in Kaggle website has been considered in this paper. The dataset contains 27,455 images (size 28 * 28) of hand gestures made by different people. Deep learning technique is used based on CNN which automatically learns and extracts features for classifying each gesture. The paper does comparative study with four recent works. The proposed model reports 99% test accuracy.

Prediction, forecast, and prognosis about the future is an interesting analysis and one of the most popularly used applications in various domains of science and technology, management science, medical science, agricultural science, and many more applied fields. This task of prediction can be automated using classification techniques, a part of supervised machine learning (ML) algorithms. The classification technique is applied to classify true/false outcome or multiple classes of any prediction. Logistic regression, Naive Bayes, k-nearest neighbor, decision trees, support vector machines, and artificial neural network are some of the popularly used classification algorithms in various domains. Depending on the case studies from various domains, clean and considerable amount of data is one of the basic requirements for class prediction. Some of the case studies from various domains are discussed in this chapter. The datasets, classification programs written in Python are provided for each case study. The algorithms to be used vary depending upon various factors such as binary or multiclass classification, multivariate or temporal datasets, size of training data, fast or accurate prediction. The ML is an upcoming field, the Python language is very popular for building software, with its rich set of libraries for efficient data management such as pandas, scientific numerical computation using NumPy, ML algorithms based on scikit-learn, visualization tools using Matplotlib, neural network algorithms using Keras, etc. The author in this chapter will showcase various ML tasks with case studies using Python programing language so that the novice readers in this field can start their learning curve smoothly.

Recommender systems provide suggestions to appropriate recipients based on suggestions and inputs given by the clients or users. The system guides the user in a tailored way to objects of interest in a larger space of possible options. In this paper, with the help of a case study the analysis of the results are obtained for different algorithms used for recommender systems—text based, semantic text based, visual similarity based product similarity. Correctness measure is performed using mean reciprocal rank metric. The paper concludes that no single algorithm is suitable for all recommender systems. Based on the type of user preferences, ensemble and hybrid algorithms are the future for recommendation.

With an alarming increase in pollution level in various parts of the world, the study of air pollution monitoring has become one of the most demanding areas in national and international arena. The paper makes an attempt to serve as a concise guide in providing basic understanding of types of remote sensing, sources of satellite data, geospatial data analysis, aerosol optical depth for air pollution monitoring. The paper gives important references on the datasets and available formats, prerequisite tools required for processing and analysis of the data. The futuristic insights finally discuss the role of machine learning in correcting AOD values as well as filling gaps for missing aerosol values when compared to ground based sensors.