Machine Learning for Big Data Analyis

A metaheuristic is a higher-level procedure designed to select a heuristic (partial search algorithm) that may lead to a sufficiently good solution to an optimization problem, especially with incomplete or imperfect information. The basic principle of metaheuristics is to sample a set of solutions which is large enough to be completely sampled. As metaheuristics make few assumptions about the optimization problem to be solved, they may be put to use in a variety of problems. Metaheuristics do not however, guarantee that a globally optimal solution can be found on some class of problems since most of them implement some form of stochastic optimization. Hence the solution found is often dependent on the set of random variables generated. By searching over a large set of feasible solutions, metaheuristics can often find good solutions with less computational effort than optimization algorithms, iterative methods, or simple heuristics. As such, they are useful approaches for optimization problems.

Even though the metaheuristics are robust enough to yield optimum solutions, yet they often suffer from time complexity and degenerate solutions. In an effort to alleviate these problems, scientists and researchers have come up with the hybridization of the different metaheuristic approaches by conjoining with other soft computing tools and techniques to yield failsafe solutions. In a recent advancement, quantum mechanical principles are being employed to cut down the time complexity of the metaheuristic approaches to a great extent. Thus, the hybrid metaheuristic approaches have come a long way in dealing with the real life optimization problems quite successfully.

Proper and faithful analysis of digital images has been in the helm of affairs in the computer vision research community given the varied amount of uncertainty inherent in digital images. Images exhibit varied uncertainty and ambiguity of information and hence understanding an image scene is far from being a general procedure. The situation becomes even graver when the images become corrupt with noise artifacts. The applications of proper analysis of images encompass a wide range of applications which include image processing, image mining, image inpainting, video surveillance, intelligent transportation systems to name a few. One of the notable areas of research in image analysis is the estimation of age progression in human beings through analysis of wrinkles in face images, which can be further utilized for tracing unknown or missing persons. Hurdle detection is one of the common tasks in robotic vision that have been done through image processing, by identifying different type of objects in the image and then calculating the distance between robot and hurdles. Image analysis has a lot to contribute in this direction.

Processing of color images takes the problem of image analysis to a new dimension. Apart from processing and analysis of the color gamut which involves a lot of computational overhead, the problem also involves analysis of the varied amount of uncertainty exhibited by the color images. A video is a very fast movement of pictures. Video analysis as a part of image analysis focuses on Shot Boundary Detection (SBD), dissolve detection, detection of gradual transitions and detection of fade ins/outs.

Recent trends in research on image analysis rely heavily on pose and gesture analysis. Typical applications include human-machine interaction, behavior analysis, video surveillance, annotation, search and retrieval, motion capture for the entertainment industry and interactive web-based applications. Real-time video analysis algorithms mainly focus on hand and head tracking and gesture analysis. A faithful gesture recognition algorithm can be implemented with techniques borrowed from computer vision and image processing.

The evolution of the functional Magnetic Resonance Imaging (fMRI) has led to proper analysis of the study mechanisms in the brain. Several statistical machine learning algorithms have been developed to analyze fMRI data in order to learn to identify and track the cognitive processes that give rise to observed fMRI data. These fMRI data can be decoded faithfully using image analysis algorithms to understand as to which word a person is thinking about based only on the neural activity captured in their fMRI data.

Remote sensing is the acquisition of information about an object or phenomenon without making physical contact with the object and thus in contrast to on site observation. Remote sensing is a sub-field of geography. In modern usage, the term generally refers to the use of aerial sensor technologies to detect and classify objects on Earth (both on the surface, and in the atmosphere and oceans) by means of propagated signals (e.g. electromagnetic radiation). It may be split into active remote sensing (when a signal is first emitted from aircraft or satellites) or passive (e.g. sunlight) when information is merely recorded. Proper analysis of remotely sensed images plays a key role in understanding the physical features of a terrain and its associated surrounding.

Several books have been published in this direction of image analysis using metaheuristic approaches. Most of them have only dealt with the subject matter using the traditional metaheuristic techniques. But none of them addresses the research issues using hybrid intelligent approaches. The proposed book is targeted to bring forward the latest research in image analysis using the hybrid intelligent metaheuristic paradigms. The use of the hybridization of the metaheuristic approaches will definitely carve out robust and fail-safe solutions for future generation image processing systems. It will entice the readers to develop as well as foresee suitable solutions for the ever evolving problems in image analysis and understanding