Face Verification Resistant to Spoofing based on Lib Movements

Khanehgir, Saeed | 2021

16 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54128 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Ghaemmaghami, Shahrokh; Eghlidos, Taraneh
  7. Abstract:
  8. Identity verification is a key part of identity reidentification process. Nowadays, Identity reidentification using face-based algorithms are popular in learning and vision area due to their generality and accessibility of this body organ. Using a fake image, occlusions on face and appearance changes like makeup can cause distortion in face verification systems which can be a drop in function of such systems. Most of these face verification models like DeepFace, FaceNet, ArcFace and SphereFace use convolution networks as their major architecture. These models, in addition to their large storage consuming and high computational complexity, due to using face as their major feature, are not robust against aforementioned attacks and their accuracy drops impressively. Therefore, we decided to introduce a face verification system that is independent of the individual's face and uses features in the face that are resistant to these attacks. The proposed system in this study uses lip changes as a feature. The system uses a combination of two architecture, a variational autoencoder architecture and a long short-term memory network. The reason for using this architecture is that the combination of the above two networks is able to learn the probabilistic distribution of its input, which is usually sequential. Therefore, we will be able to learn the distribution of lip changes in each person and perform the face verification operation based on the obtained distribution parameters. Since the pattern of lip changes is different in each person and the attacks mentioned above do not affect this organ, we were able to obtain a model with high accuracy and low volume by using a suitable learning method. The proposed method is examined using VidTIMIT database and EER evaluation metric. This system with good error rate of 1.52% and AUC of 96.8% shows good performance
  9. Keywords:
  10. Face Verification ; Long Short Term Memory (LSTM) ; Deep Learning ; Variational Autoencoder ; Face Spoofing ; Identity Verification

 Digital Object List