Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

PCA vs LDA

4,090 views

Published on

In context of Face recognition

Published in: Engineering
  • Dating direct: ❶❶❶ http://bit.ly/39sFWPG ❶❶❶
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Sex in your area is here: ♥♥♥ http://bit.ly/39sFWPG ♥♥♥
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

PCA vs LDA

  1. 1. Comparison between PCA and LDA
  2. 2.  PCA applied to data identifies the directions in the feature space (principal components) that account for the most variance in the data.  Linear Discriminant Analysis (LDA) tries to identify characteristics that account for the most variance between classes.  Eigenfaces (PCA)  project faces onto a lower dimensional sub-space  no distinction between inter- and intra-class variabilities.  optimal for representation but not for discrimination.  Fisherfaces (LDA)  find a sub-space which maximizes the ratio of inter-class and intra-class variability.  same intra-class variability for all classes
  3. 3. There has been a tendency in the computer vision community to prefer LDA over PCA. This is mainly because LDA deals directly with discrimination between classes while PCA does not pay attention to the underlying class structure.  LDA has lower error rates LDA works well even if different illumination LDA works well even if different facial express.
  4. 4. 4 Is LDA always better than PCA? This paper shows that when the training set is small, PCA can outperform LDA. A. Martinez, A. Kak, "PCA versus LDA", IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 23, no. 2, pp. 228-233, 2001 Our project also show the same result. When the number of samples is large and representative for each class, LDA outperforms PCA.
  5. 5. Implementation  Training Database  There are five classes (person)  Each class has Eight elements (face images)  1 to 8 in class 1, 9 to 16 in class 2,and so on…  Test faces  10 test images  Result : PCA and show different face images even using the same training database
  6. 6.  Efficiency=(Number of correct recognition)/(Total number of test faces)

×