Summary of Paint Outside the Box: Synthesizing and Selecting Training Data For Visual Grounding, by Zilin Du et al.
Paint Outside the Box: Synthesizing and Selecting Training Data for Visual Grounding
by Zilin Du, Haoxin Li, Jianfei Yu, Boyang Li
First submitted to arxiv on: 1 Dec 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 explores visual grounding, a technique that localizes image regions based on textual queries, under data-scarce settings. To address this challenge, the authors propose a novel framework called POBF (Paint Outside the Box, then Filter). This framework synthesizes images by inpainting outside the box and uses an innovative filtering scheme to identify effective training data. The scheme combines hardness scores, overfitting scores, and penalty terms to balance performance and robustness. Experimental results show that POBF outperforms leading baselines by 2.29% to 3.85% in accuracy, achieving an average improvement of 5.83%. Additionally, the authors validate the robustness and generalizability of POBF across various generative models, data ratios, and model architectures. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps computers better understand what’s happening in pictures when we ask them questions about those images. It’s like asking Siri to find a specific cat picture on your phone. The authors found that using their new method called POBF makes it work really well even when there isn’t much training data. They tested it with different kinds of computer models and showed that it works better than other methods. |
Keywords
» Artificial intelligence » Grounding » Overfitting
