Tutorial 7
Tutorial 7 of COMP435p: Biometrics Authentication
by Qu Xiaofeng, csxfqu@comp.polyu.edu.hk
- Tutorial Slides in COMP (Backup link hosted in GitHub)
- Word version of this tutorial
- Lecture Note 8
- Lecture Note 5 (mentioned in the question 4)
1. Answer the following questions
-
(a) Understand why use face (P8: 3-4) and face recognition system (P8: 5-10).
-
(b) Please point out some problems of face recognition. Can you give their solutions? (P8: 11-16)
-
(c) What are the general 3 steps of face recognition? (P8:17)
-
(d) There are two kinds of the methods in face detection and location, which are statistics-based and knowledge-based method. What difference between them? (P8: 18-22)
-
(e) There are five methods of face feature extraction (P8: 24-41). Please give their basic ideas.
-
(f) What are the advantages and disadvantages of face recognition? (P8: 48)
2. Understand the eigenfaces algorithm (P8: 26-29) and eigenfaces recognition algorithm (P8: 30-31).
3. Eigenface is PCA-based method which is introduced in Lecture 5. At the fifth step of PCA method (P5:23-24) the k most principal components are selected based on the ratio (γλ) of the eigenvalue sum of selected components to the total sum. Please decide the value of k when the threshold of γλ is 95%. (<font color=darkred>7</font>)
In P8:31 there is a training set including 27 facial images to get eigenfaces. Can you give the value of k according to the training result (eigenfaces)? (<font color=darkred>9</font>)
4. P8:34 shows some examples of local features. Please try to define the geometry features based on the woman’s face image (refer to P8:34 ) and give the number of features. (<font color=darkred>21</font>)
