Semantic coding is used to provide image representation able to make semantics emerge and to learn it. This process is based on a basic decomposition of the image content into "visual words"; using spatial relationships, these visual words are aggregated to form structures and (inter)active learning is used to associate semantics to the emerging structures. A complete OTB tool has been developed to illustrate the whole process; it allows feature extraction and "visual words" production, as well as LDA and SVM learning approachs.