Image Semantic Relation Generation
This work addresses a bottleneck in scene graph generation for applications like image retrieval and autonomous vehicles, though it is incremental as it builds on existing segmentation and text generation methods.
The paper tackles the high labor cost of constructing scene graph annotations by proposing a decoupled approach for image semantic relation generation, achieving 31 points on the OpenPSG dataset and outperforming baselines by up to 16 points.
Scene graphs provide structured semantic understanding beyond images. For downstream tasks, such as image retrieval, visual question answering, visual relationship detection, and even autonomous vehicle technology, scene graphs can not only distil complex image information but also correct the bias of visual models using semantic-level relations, which has broad application prospects. However, the heavy labour cost of constructing graph annotations may hinder the application of PSG in practical scenarios. Inspired by the observation that people usually identify the subject and object first and then determine the relationship between them, we proposed to decouple the scene graphs generation task into two sub-tasks: 1) an image segmentation task to pick up the qualified objects. 2) a restricted auto-regressive text generation task to generate the relation between given objects. Therefore, in this work, we introduce image semantic relation generation (ISRG), a simple but effective image-to-text model, which achieved 31 points on the OpenPSG dataset and outperforms strong baselines respectively by 16 points (ResNet-50) and 5 points (CLIP).