Summary of Region Of Interest Loss For Anonymizing Learned Image Compression, by Christoph Liebender et al.
Region of Interest Loss for Anonymizing Learned Image Compression
by Christoph Liebender, Ranulfo Bezerra, Kazunori Ohno, Satoshi Tadokoro
First submitted to arxiv on: 9 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Cryptography and Security (cs.CR); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
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 presents a novel approach to anonymizing sensitive data in public spaces by using customized loss functions on region of interests (ROI). The method trains an end-to-end optimized autoencoder for learned image compression that utilizes the flexibility of the learned analysis and reconstruction transforms for the task of mutating parts of the compression result. This approach enables compression and anonymization in one step on the capture device, rather than transmitting sensitive data over the network. The paper evaluates the impact of this anonymization on pre-trained foundation models for detecting faces (MTCNN) and humans (YOLOv8), considering compression rate and latency. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In public spaces, AI is used to detect and recognize people, collecting lots of personal data. But what if we don’t need to know who the person is? Just their position in the scene would be enough. The paper shows how to anonymize this data by changing the way images are compressed. This can be done on the camera itself, so sensitive data isn’t sent over the network. The results show that pre-trained models for face and human detection still work well with this anonymized data. |
Keywords
» Artificial intelligence » Autoencoder
