Towards Open-vocabulary Scene Graph Generation with Prompt-based Finetuning
This addresses a practical limitation in scene graph generation for computer vision applications, though it is incremental as it builds on existing methods with prompt-based techniques.
The paper tackles the limitation of closed-vocabulary scene graph generation by introducing an open-vocabulary setting, where models infer relations for unseen object classes, and achieves significant performance improvements over strong baselines on benchmark datasets like Visual Genome, GQA, and Open-Image.
Scene graph generation (SGG) is a fundamental task aimed at detecting visual relations between objects in an image. The prevailing SGG methods require all object classes to be given in the training set. Such a closed setting limits the practical application of SGG. In this paper, we introduce open-vocabulary scene graph generation, a novel, realistic and challenging setting in which a model is trained on a set of base object classes but is required to infer relations for unseen target object classes. To this end, we propose a two-step method that firstly pre-trains on large amounts of coarse-grained region-caption data and then leverages two prompt-based techniques to finetune the pre-trained model without updating its parameters. Moreover, our method can support inference over completely unseen object classes, which existing methods are incapable of handling. On extensive experiments on three benchmark datasets, Visual Genome, GQA, and Open-Image, our method significantly outperforms recent, strong SGG methods on the setting of Ov-SGG, as well as on the conventional closed SGG.