Summary of Synmorph: Generating Synthetic Face Morphing Dataset with Mated Samples, by Haoyu Zhang et al.
SynMorph: Generating Synthetic Face Morphing Dataset with Mated Samples
by Haoyu Zhang, Raghavendra Ramachandra, Kiran Raja, Christoph Busch
First submitted to arxiv on: 9 Sep 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 method generates a large-scale, public-available synthetic face morphing dataset to overcome the limitations of existing datasets in the field of face recognition systems. The dataset comprises 2450 identities and over 100k morphs, showcasing high-quality samples and varying morphing algorithms for single and differential attack detection. Face image quality assessment and vulnerability analysis are used to evaluate the dataset’s effectiveness, with results benchmarked against a state-of-the-art (SOTA) synthetic dataset and a non-synthetic counterpart. The proposed dataset shows improvement over the SOTA in terms of biometric sample quality and morphing attack potential. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to create fake face images is developed to help protect people’s privacy. This method generates a large collection of synthetic (fake) face images with different identities and morphing styles, which can be used to train computers to detect when someone is trying to manipulate their facial features. The dataset has over 100,000 images and is designed for use in both single and combined attacks on face recognition systems. The quality of the generated faces is evaluated and compared to other datasets, showing that it’s a better way to test how well these systems work. |
Keywords
» Artificial intelligence » Face recognition
