A Multi Criteria Decision Making Based Approach for Semantic Image Annotation

This paper presents a novel approach for automatic image annotation that integrates visual topics and regional contexts, addressing the limitations of previous methods that overlooked the relationships between image regions. The proposed model uses multi-criteria decision-making (MCDM) based on the weighted sum method (WSM) to combine these two types of information, significantly improving annotation accuracy as demonstrated by experiments on the 5k Corel dataset. The approach also extends traditional statistical models to better capture the semantic relationships in images, enhancing the effectiveness of image retrieval.