Fuzzy Knowledge-based Image Annotation Refinement

Automatic image annotation methods automatically assign labels to images in order to facilitate tasks such as image retrieval. Incorrect labels may negatively influence the search results so image annotation should be as accurate as possible. Labels pertaining to objects or to whole scenes are commonly used for image annotation, and precision is especially important in case when scene labels are inferred from objects, as errors in the object labels may propagate to the scene level. To improve the annotation precision, the idea is to infer which labels are incorrect using the context of other labels and the knowledge about objects and their relations. This procedure is here referred to as annotation refinement. The proposed approach used in this paper includes a fuzzy knowledge base and uses the fuzzy inference algorithms to detect and discard automatically obtained object labels that do not fit the context of other detected objects.