Summary of Negative Prototypes Guided Contrastive Learning For Wsod, by Yu Zhang et al.
Negative Prototypes Guided Contrastive Learning for WSOD
by Yu Zhang, Chuang Zhu, Guoqing Yang, Siqi Chen
First submitted to arxiv on: 4 Jun 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 proposes a novel approach to weakly supervised object detection, leveraging image-level annotations to detect objects. The authors introduce Negative Prototypes Guided Contrastive learning (NPGC) architecture, which incorporates negative prototypes, online-updated global feature banks, and pseudo-label sampling modules. This framework optimizes proposal features by attracting similar class samples and pushing away different class samples in the embedding space. The method achieves state-of-the-art performance on VOC07 and VOC12 datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The researchers developed a new way to detect objects using only image labels. They created a special kind of prototype, called negative prototypes, which are proposals that are misclassified for a category not present in the label. This helps use the weak label more effectively. The team also built an online feature bank that stores both positive and negative prototypes. To make sure the model is learning from reliable instances, they designed a module to sample pseudo labels based on feature similarity with corresponding prototypes. Finally, they used contrastive learning to optimize proposal features by moving similar class samples closer together and pushing away different class samples. |
Keywords
» Artificial intelligence » Embedding space » Object detection » Supervised
