Summary of Anople: Few-shot Anomaly Detection Via Bi-directional Prompt Learning with Only Normal Samples, by Yujin Lee et al.
AnoPLe: Few-Shot Anomaly Detection via Bi-directional Prompt Learning with Only Normal Samples
by Yujin Lee, Seoyoon Jang, Hyunsoo Yoon
First submitted to arxiv on: 24 Aug 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 The paper introduces AnoPLe, a multi-modal prompt learning method for few-shot anomaly detection (FAD) without prior knowledge of anomalies. It simulates anomalies and uses bidirectional coupling to facilitate deep interaction between textual and visual prompts. Additionally, it integrates a lightweight decoder with a learnable multi-view signal trained on multi-scale images to enhance local semantic comprehension. The paper achieves strong FAD performance, recording 94.1% and 86.2% Image AUROC on MVTec-AD and VisA respectively, with only around a 1% gap compared to the state-of-the-art (SoTA), despite not being exposed to true anomalies. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary AnoPLe is a new way to find things that don’t belong in a picture. It looks at both text and images together to learn what’s normal and what’s not. This helps it detect unusual things even when there are very few examples to learn from. The team tested AnoPLe on two big datasets and found that it worked really well, almost as good as the best other methods. But the best part is that AnoPLe doesn’t need to see any actual weird or abnormal pictures to do its job. |
Keywords
» Artificial intelligence » Anomaly detection » Decoder » Few shot » Multi modal » Prompt
