دانلود کتاب تلفیق اطلاعات – روش های یادگیری ماشینی

Thanks to the quick development of multimedia techniques, a large amount of data can be collected in different ways. In a large number of real-world applications, the source data or signal is obtained from multiple domains or represented by different measurements. For example, a person could be verified via fingerprint, palmprint, or iris, and a pair of given samples can be measured by the Euclidean or Mahalanobis distance. Generally, the unique representation fails to properly extract valuable information for various applications. Thus, it is significant to conduct research on how to efficiently achieve information fusion. Information fusion is to combine different types of data to get more accurate, consistent, and useful information compared with any individual modality. Based on the fusion stage, information fusion processes are usually classified into three categories, including low, intermediate, and high levels.