In recent years, social media has become an essential part of networking and communication environment online. As people use this environment to connect one another and retrieve different information of their interest, one of the most popular information which is being shared through the social media network is multimedia data such as photos, video and music. Along with this popularity, various multimedia-based applications are developed to entertain the users. These include art creation (e.g. applications for a portrait photo into a pencil drawing or a painting) and cultural creation (e.g. applications for processing a scenery photo into a Japanese wood-print art. In the field of multimedia image processing, semantic multi-media retrieval has been in a main focus of research for the last two decades. Recently a large number of cutting-edge algorithms have been developed using object recognition techniques for predetermined domains such as identifying animals and plants over the Internet or large datasets. In order to overcome the so-called "semantic gap", content-based image retrieval methods using low-level visual features have been exploited and applied at early years until now. Today newer methods have been developed. Artificial intelligence and machine learning techniques applying mainly unsupervised learning is a focused research area. A more recent research, which is gaining an increased interest, is connecting low-level features with contextual and conceptual information. Meanwhile, emergence of high-end devices such as smart phones comes with a wide variety of metadata. Selectively applying some useful metadata can further improve intelligence of the semantic multi-media research environment.