Zero-Shot Object Classification With Large-Scale Knowledge Graph

Kohei Shiba, Yusuke Mukuta, Tatsuya Harada; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2023, pp. 4991-4998

Zero-shot learning is used to predict unseen categories and can solve problems such as dealing with unseen categories that were not anticipated at the time of training and the lack of labeled datasets. One method for zero-shot object classification is to use a knowledge graph, which is a set of explicit knowledge. Because recognition is limited to the categories contained in the knowledge graph, and the relationships among categories are expected to be quantitatively and qualitatively richer depending on the graph size, it is desirable to handle a large-scale knowledge graph that contains as many categories as possible.We used a knowledge graph that contains approximately seven times as many categories as the knowledge graphs used mainly in existing research to enable the classification of a larger number of categories and to achieve more accurate recognition.When using a large-scale knowledge graph, the number of noisy nodes and edges is expected to increase.Therefore, we propose a method to extract useful information from the entire graph using positional relationships between categories and types of edges in the knowledge graph. We classify images that were earlier unclassifiable in existing research and show that the proposed data extraction method improves performance compared to using the entire graph.