Open Jingkang50 opened 2 years ago
Thanks for sharing the information and your awesome dataset. We find there is a typo when retrieving the information from Table 1 of 'Deep Image Synthesis from Intuitive User Input: A Review and Perspectives'. COCO-Stuff should be ticked instead of COCO, we will correct it in the table and also include your dataset in the table.
Cool, that makes sense. Thank you Dr. Zhan for your clarification!
Thank you for your nice survey! It really gives good guidance for the community. However, we notice that in Table 1, the modality of scene graph in COCO is ticked. May I ask where is the COCO scene graph comes from? Actually, our recent work PSG provides a high-quality scene graph annotation based on COCO panoptic segmentation, so that
Semantic Map, Bounding Box, Text, Scene Graph
are all available in the PSG dataset.You could get the general idea of PSG dataset by glancing our codebase or interactively explore the dataset on our website. We would really appreciate if the PSG dataset could be included in the table and contribute to the image editing community.