Summary of Siavc: Semi-supervised Framework For Industrial Accident Video Classification, by Zuoyong Li et al.
SIAVC: Semi-Supervised Framework for Industrial Accident Video Classification
by Zuoyong Li, Qinghua Lin, Haoyi Fan, Tiesong Zhao, David Zhang
First submitted to arxiv on: 23 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 semi-supervised learning method called SIAVC for industrial accident video classification. The method incorporates two key components: the Super Augmentation Block (SAB), which adds Gaussian noise and masks video frames based on historical loss, and the Video Cross-set Augmentation Module (VCAM), which generates diverse pseudo-label samples from high-confidence unlabeled data. The authors also introduce a new industrial accident surveillance video dataset, ECA9, with frame-level annotation, to evaluate their proposed method. Compared to state-of-the-art semi-supervised learning methods, SIAVC achieves excellent performance on both the ECA9 and Fire Detection datasets, with accuracy rates of 88.76% and 89.13%, respectively. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us learn better from videos without needing a lot of labeled information. The authors created a new way to improve video classification by using unlabeled data. They designed two special tools: one adds noise to the videos, and another generates fake labels for the unlabeled parts. This method is good at classifying industrial accident videos and can even detect fires. The researchers also made a big dataset with labeled frames to test their method. |
Keywords
» Artificial intelligence » Classification » Semi supervised
