Summary of Reconstruction-based Anomaly Localization Via Knowledge-informed Self-training, by Cheng Qian et al.
Reconstruction-Based Anomaly Localization via Knowledge-Informed Self-Training
by Cheng Qian, Xiaoxian Lao, Chunguang Li
First submitted to arxiv on: 22 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 The proposed Knowledge-Informed Self-Training (KIST) method integrates domain expert knowledge into a reconstruction model for improved anomaly localization in images. By utilizing weakly labeled anomalous samples, KIST generates pixel-level pseudo-labels and incorporates them into a novel loss function that promotes normal pixel reconstruction while suppressing anomalous pixels. This approach outperforms existing reconstruction-based methods on various datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary KIST is a new way to find unusual things in pictures using information from experts. It takes weakly labeled abnormal samples and uses this knowledge to create fake labels for each pixel. Then, it uses these fake labels to make the model better at finding normal pixels while ignoring abnormal ones. This helps improve image anomaly localization. |
Keywords
* Artificial intelligence * Loss function * Self training
