Indexing of large-scale multimedia signals

The effective and efficient indexing of multimedia signals has attracted extensive research interests over last decades. Nowadays, indexing is not merely a technology that helps multimedia search but also the basis of a wide variety of applications and services, such as recommendation, advertising, and personalization. On the other hand, we have witnessed the explosive growing of multimedia data. Such scale brings significant challenges and profound impacts to both the text-based indexing and feature-level indexing of multimedia signals. For example, it is very challenging for many annotation and semantic hashing algorithms to effectively and efficiently handle large-scale multimedia signals, especially when the scale comes up from tens of thousands to tens of millions or even billions. Fortunately, along with the growth of multimedia signals, more and more resources also become available, such as the associated metadata, context and social information. In addition, collaborative tagging, a representative behavior of web 2.0, enables the availability of tags for a large amount of multimedia signals on the Internet. These facts have provided opportunities to tackle the difficulties in large-scale multimedia indexing. This special issue is organized with the purpose of introducing novel research work on indexing of large-scale multimedia signals. Submissions have come from an open call for paper. With the assistance of professional referees, 24 papers are selected after at least two rounds of rigorous reviews. These papers cover widely subtopics of largescale multimedia indexing, including multimedia annotation, multimedia hashing, visual search, copy detection, and so on. We divide the whole special issue into 5 parts according to the themes of the papers. The first and also the largest part contains 11 papers that are related to multimedia classification or annotation. This indicates that multimedia classification and annotation is still an extremely active research field. The first paper, “Image Classification Using Harr-Like Transformation of Local Features with Coding Residuals”, applies Harr-like transformation on local features to combine the spatial information as well as the correlations of local features. These Harr-like transformed local features are then encoded using non-negative sparse coding for classification. In the second paper