Summary of Scene Graph Generation Strategy with Co-occurrence Knowledge and Learnable Term Frequency, by Hyeongjin Kim et al.
Scene Graph Generation Strategy with Co-occurrence Knowledge and Learnable Term Frequency
by Hyeongjin Kim, Sangwon Kim, Dasom Ahn, Jong Taek Lee, Byoung Chul Ko
First submitted to arxiv on: 21 May 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel approach to Scene Graph Generation (SGG), which represents the relationships between objects in an image as a graph structure. The authors address two key limitations of previous SGG studies: failing to reflect co-occurrence of objects and only addressing the long-tail problem from the perspectives of sampling and learning methods. To overcome these challenges, they introduce CooK, which incorporates Co-occurrence Knowledge between objects, and TF-l-IDF to tackle the long-tail issue. The proposed model outperforms existing state-of-the-art models in the SGGen subtask by up to 3.8% on the SGG benchmark dataset, demonstrating generalization ability and uniform performance improvement for all MPNN models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us understand images better by creating a special graph that shows how objects relate to each other. Right now, this is hard to do because previous methods didn’t take into account when different objects appear together in an image. They also had trouble with rare or unusual object combinations. The authors came up with a new way to make these relationships work better called CooK. This helps their model learn and improve over time. When they tested this new method, it did much better than the previous best models. |
Keywords
» Artificial intelligence » Generalization
