LAUSR

Image annotation is the process of assigning metadata to images, allowing effective retrieval by text-based search techniques. Despite the lots of efforts in automatic multimedia analysis, automatic semantic annotation of multimedia is still inefficient due to the problems in modeling high-level semantic terms. In this paper, we examine the factors affecting the quality of annotations collected through crowdsourcing platforms. An image dataset was manually annotated utilizing: (1) a vocabulary consists of preselected set of keywords, (2) an hierarchical vocabulary and (3) free keywords. The results show that the annotation quality is affected by the image content itself and the used lexicon. As we expected while annotation using the hierarchical vocabulary is more representative, the use of free keywords leads to increased invalid annotation. Finally, it is shown that images requiring annotations that are not directly related to their content (i.e., annotation using abstract concepts) lead to accrue annotator inconsistency revealing in that way the difficulty in annotating such kind of images is not limited to automatic annotation, but it is a generic problem of annotation.