II-20: Intelligent and pragmatic analytic categorization of image collections

In this paper, we introduce 11-20 (Image Insight 2020), a multimedia analytics approach for analytic categorization of image collections. Advanced visualizations collections exist, but they need tight integration with machine model to support the task categorization. Directly employing computer vision and interactive learning techniques gravitates towards search. Analytic categorization, however, is not classification (the difference between two called pragmatic gap): human adds/redefines/deletes categories relevance on fly build insight, whereas classifier rigid non-adaptive. that truly brings user insight requires flexible allows dynamic sliding exploration-search axis, as well semantic interactions: thinks about data mostly in terms. three major contributions closing gap. Firstly, new closely follows user's interactions dynamically models her relevance. II-20's model, addition matching exceeding state art's ability produce relevant suggestions, slide axis without any additional input from side. Secondly, dynamic, 1-image-at-a-time Tetris metaphor synergizes model. It well-trained analyze collection by itself minimal interaction complements classic grid metaphor. Thirdly, fast-forward interaction, allowing harness quickly expand (“fast-forward”) relevance, expands dictionary. Automated experiments show outperforms existing art also demonstrate metaphor's quality. User studies further confirm II-20 an intuitive, efficient, effective tool.