Automatic Generation of Multimodal Learning Objects from Electronic Textbooks

Abstract

Technology-supported learning systems (TSLS) (e-learning, educational software, electronic books) require an appropriate representation of the knowledge to be learned that is the domain module. Domain module enables to guide the students in the learning process or to learn by themselves. The authoring of the domain module is cost- and labor intensive. Achieving greater accuracy in the generation of domain module is an issue. The proposed system considers the source of electronic textbooks that are documents in different formats—pdf, doc as input. The system is designed to efficiently handle the multimodal learning objects (images and captions) that are present in electronic textbooks. Also, the proposed system is designed to improve the generation of learning domain ontology (LDO), which employs ontology construction using natural language processing (NLP) techniques. Thus, improvisation in LDO guarantees to achieve a higher accuracy.