Summary of Glotsformer: Global Video Text Spotting Transformer, by Han Wang and Yanjie Wang and Yang Li and Can Huang
GloTSFormer: Global Video Text Spotting Transformer
by Han Wang, Yanjie Wang, Yang Li, Can Huang
First submitted to arxiv on: 8 Jan 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 Video Text Spotting (VTS) is a crucial computer vision task that predicts text trajectories and content in videos. Traditional approaches focus on local associations, ignoring temporal information and morphological characteristics. This paper introduces the Global Video Text Spotting Transformer (GloTSFormer), which models VTS as global associations and utilizes the Gaussian Wasserstein distance to correlate frames. The GloTSFormer achieves significant improvements over state-of-the-art methods on public datasets, including a 4.6 MOTA increase on ICDAR2015 with an absolute improvement of 56.0 MOTA. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine trying to find words in a moving video. This is called Video Text Spotting (VTS). Before this paper, people focused on finding small parts of text and didn’t use the whole video or how the text changes over time. The researchers created a new way to do VTS using a special kind of computer program called Transformer. It works by looking at multiple frames in the video at once and using distance to find the best match. This new method, called GloTSFormer, does much better than previous methods on big datasets. |
Keywords
» Artificial intelligence » Transformer
