Manzanilla: An Image Annotation Tool for TKB Building

Manzanilla: An Image Annotation Tool for TKB Building Arianne Reimerink, Pilar León-Araúz University of Granada Department of Translation and Interpreting, C/ Buensuceso, 11, 18008 Granada (Spain) arianne@ugr.es, pleon@ugr.es Abstract Much has been written regarding the importance of combining visual and textual information to enhance knowledge acquisition (Paivio, 1971, 1986; Mayer & Anderson, 1992). However, the combination of images and text still needs further analysis (Faber, 2012; Prieto, 2008; Prieto & Faber, 2012). An in-depth analysis of the features of images provides the means to develop selection criteria for specific representation purposes. The combination of conceptual content, image type based on morphological characteristics, and functional criteria can be used to enhance the selection and annotation of images that explicitly focus on the conceptual propositions that best define concepts in a knowledge base. Manzanilla is an image annotation tool specifically created for EcoLexicon, a multilingual and multimodal terminological knowledge base (TKB) on the environment. It is powered by Camomile (Poignant et al., 2016) according to the selection and annotation criteria resulting from ten years of research on multimodality within the framework of Frame-Based Terminology (FBT; Faber, León-Araúz & Reimerink, 2014). The tool was created to enhance the consistency of knowledge representation through images with the conceptual knowledge in EcoLexicon and to improve image reusability. Keywords: image annotation, multimodal knowledge representation, EcoLexicon, Camomile 1. Introduction Manzanilla is an image annotation tool specifically created for EcoLexicon, a multilingual and multimodal terminological knowledge base (TKB) on the environment1. It was developed with Camomile (Collaborative Annotation of multi-MOdal, multI-Lingual and multi-mEdia documents; Poignant et al., 2016)2 according to the selection and annotation criteria resulting from ten years of research on multimodality within the framework of Frame-Based Terminology (FBT; Faber, León-Araúz & Reimerink, 2014). The tool was created to enhance the consistency of knowledge representation through images with the conceptual knowledge in EcoLexicon and to improve image reusability. Currently, images are stored in the TKB in association with concept entries according to the semantic content described in their definition, and are thus regarded as a whole and only linked to the concept itself. Other knowledge bases, such as BabelNet, the automatically constructed multilingual encyclopedic dictionary and semantic network (Navigli & Ponzetto, 2012), also uses this approach. However, regarding images as a whole does not allow for a more fine-grained annotation where the semantic relations between different concepts represented in an image are made explicit. Our new approach is that images should not be stored in the TKB as the representation of a concept, but as the representation of a set of conceptual propositions (concept-relation-concept triples) more in line with the conceptual structure of EcoLexicon. Therefore, images must be annotated according to semantic and morphological information and stored in a separate repository. Since each image activates several propositions and each proposition can be activated by different concepts, one image can then be linked to several concept entries. This would enhance the reusability of images, improve the consistency of the TKB and avoid duplicating workload (Reimerink, León-Araúz & Faber, …

Content truncated to 500 words