Summary of Whispernetv2: Slowfast Siamese Network For Lip-based Biometrics, by Abdollah Zakeri et al.
WhisperNetV2: SlowFast Siamese Network For Lip-Based Biometrics
by Abdollah Zakeri, Hamid Hassanpour, Mohammad Hossein Khosravi, Amir Masoud Nourollah
First submitted to arxiv on: 11 Jul 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 A novel lip-based biometric authentication network called WhisperNetV2 has been proposed to improve the accuracy of this technique. The existing research on lip-based biometric authentication (LBBA) has ignored the potential emotional influence of clients during video acquisition, which can impact facial expressions and speech tempo. To address this limitation, the proposed WhisperNetV2 network leverages a deep Siamese structure with triplet loss and three identical SlowFast networks as embedding networks. The SlowFast network is well-suited for this task due to its ability to extract motion-related features (behavioral lip movements) and visual features (physiological lip appearance). The network was trained using the CREMA-D dataset, achieving an Equal Error Rate (EER) of 0.005 on the test set. This state-of-the-art LBBA method demonstrates improved accuracy compared to similar methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to recognize people by their lips has been developed. This technique is important because it can be used for both physical and behavioral traits. The problem with this method is that it doesn’t consider how the person being recognized might be feeling when they’re recorded. This could affect how their face looks and how fast they talk. To fix this, a new network called WhisperNetV2 was designed. It uses special features from videos to recognize people’s lips better. The new network did very well on some tests, showing that it’s one of the best ways to do lip recognition. |
Keywords
» Artificial intelligence » Embedding » Triplet loss
