face recognition system using LBP

15,910 views

Published on

Part 1

20 Comments
19 Likes
Statistics
Notes
No Downloads
Views
Total views
15,910
On SlideShare
0
From Embeds
0
Number of Embeds
9
Actions
Shares
0
Downloads
0
Comments
20
Likes
19
Embeds 0
No embeds

No notes for slide

face recognition system using LBP

  1. 1. Final Year Project Face Recognition Using Local Features
  2. 2. Introduction <ul><li>- What is Face Recognition? </li></ul><ul><li>- Using Global Features: </li></ul><ul><li>Example: PCA and LDA </li></ul><ul><li>- Using Local Features: </li></ul><ul><li>Example: Gabor and LBP </li></ul><ul><li>- Goal of Face Recognition. </li></ul>
  3. 3. Introduction <ul><li>- Objectives of Research: </li></ul><ul><li>1- To Study Face Recognition Systems. </li></ul><ul><li>2- To design and develop Face Recognition Systems. </li></ul><ul><li>3- To implement Face Recognition System using enhanced local </li></ul><ul><li>features. </li></ul>
  4. 4. Literature Review: Biometrics <ul><li>- Biometrics: </li></ul>
  5. 5. Literature Review: Biometrics(2) <ul><li>- Biometrics: From Greek words. </li></ul><ul><li>- Bio= life , metrics= to measure . </li></ul><ul><li>- Biometrics : identify and verify a person based on their Psychological and behavioral characteristics. </li></ul><ul><li>- This first type of biometrics was fingerprint. </li></ul>
  6. 6. Literature Review: Face Recognition <ul><li>- Face Recognition: </li></ul><ul><li>- Verification: (1:1) </li></ul><ul><li>- Required Unique ID and Biometric Sample. </li></ul><ul><li>- It will compare the biometric template (sample) with the one it has on records . </li></ul><ul><li>- “Match” or “Not Match”. </li></ul><ul><li>- Identification: (1:M) </li></ul><ul><li>- Required Only Biometric Sample. </li></ul><ul><li>- It will compare the biometric template (sample) using a smart algorithm, with each one of the records in a file. </li></ul><ul><li>- positive ID of specific Identity given by its unique User ID. </li></ul>
  7. 7. Literature Review: Face Recognition Features
  8. 8. Literature Review: Face Recognition Features(2) <ul><li>- Face Recognition Features: </li></ul><ul><li>- Global Features: </li></ul><ul><li>- Focus on the Whole Entire Image. </li></ul><ul><li>- Less Accuracy. </li></ul><ul><li>- Local Features: </li></ul><ul><li>- Focus on the local features of the face, which help to identify and verify the persons using the unique details in the face. - More Accuracy. </li></ul>
  9. 9. Literature Review: Face Recognition Algorithms <ul><li>- Face Recognition Algorithms: </li></ul><ul><li>- PCA (Principle Component Analysis): </li></ul><ul><li>- Works on Global Features. </li></ul><ul><li>- It is the most famous. </li></ul><ul><li>- It is called Eigenface. </li></ul><ul><li>- It tries to find a lower dimensional subspace to describe the original face space. </li></ul><ul><li>- there are statistical equations will be used to get the required image </li></ul>
  10. 10. Literature Review: Face Recognition Algorithms(2) <ul><li>- ICA (Independence Component Analysis): </li></ul><ul><li>- It is a statistical signal processing technique. </li></ul><ul><li>- It is a special case of redundancy reduction technique and it represents the data in terms of statistically independent variables. </li></ul><ul><li>- Its goal is to minimize the statistical dependence between basic vectors. </li></ul><ul><li>- Provides powerful data representation than PCA. </li></ul>
  11. 11. Literature Review: Face Recognition Algorithms(3) <ul><li>- LDA (Linear Discriminate Analysis): </li></ul><ul><li>- Is a dimensionality reduction technique. </li></ul><ul><li>- It searches for those vectors in the underlying space that best discriminate among classes, </li></ul><ul><li>- Its main idea is to find a linear transformation such that feature clusters are most separable, which can be achieved through scatter matrix analysis. </li></ul>
  12. 12. Literature Review: Face Recognition Algorithms(4) <ul><li>- LBP (Local Binary Pattern): </li></ul><ul><li>- It works on Local Features </li></ul><ul><li>- LBP operator: summarizes the local special structure of an image. </li></ul><ul><li>- LBP is defined as an ordered set of binary comparisons of pixel intensities between the center pixel and its eight surrounding pixels. </li></ul>
  13. 13. Literature Review: Face Recognition Algorithms(5) <ul><li>- LBP (Local Binary Pattern): . </li></ul><ul><li>- decimal form of the resulting 8-bit word (LBP code) can be expressed as follows: </li></ul>
  14. 14. Literature Review: Face Recognition Algorithms(6) <ul><li>- LBP (Local Binary Pattern): </li></ul><ul><li>- We can also do this comparison by applying the following formula: </li></ul>
  15. 15. Methodology: LBP <ul><li>- LBP (Local Binary Pattern): </li></ul><ul><li>- It is used to determine the local features in the face. </li></ul><ul><li>- It works by using basic LBP operator. </li></ul><ul><li>- in a matrix originally of size 3×3, the values are compared by the value of the centre pixel, then binary pattern code is produced. </li></ul><ul><li>- The LBP code is obtained by converting the binary code into decimal one. </li></ul>
  16. 16. Methodology: LBP Histograms <ul><li>- LBP Histograms (Local Binary Pattern): </li></ul><ul><li>- Each pixel of an image is labeled with an LBP code . </li></ul><ul><li>- First it will divide the image to several blocks. </li></ul><ul><li>- Then it will start calculating the LBP histogram for each block. </li></ul><ul><li>- after that it will combine every LBP histogram for that image </li></ul><ul><li>- then you will get all the LBP histograms into one vector. </li></ul>
  17. 17. Methodology: LBP Histograms(2)
  18. 18. Methodology: LBP Flowchart
  19. 19. Methodology: LBP Flowchart(2) <ul><li>- LBP Process Flowchart: </li></ul><ul><li>- Capture an image then store it. </li></ul><ul><li>- The process will divide the image to several blocks. </li></ul><ul><li>- Histograms will be calculated for each block, then a histograms will be concentrated into a single vector. </li></ul><ul><li>- As a result, the facial recognition is represented by LBP and the shape of the face is obtained by concentration of different local histograms. </li></ul>
  20. 20. Proposed Solution: LBP Analysis <ul><li>- LBP Analysis: </li></ul><ul><li>- </li></ul>
  21. 21. Proposed Solution: LBP Analysis(2) <ul><li>- First, the image will be divided to several blocks; each block is as a matrix of type 3 X 3. </li></ul><ul><li>- Then, in each matrix the system will be comparing the center pixel of the block with the other pixels. </li></ul><ul><li>- If Center pixel >= other pixel = 1, else = 0. </li></ul><ul><li>- Getting the binary number of each block. </li></ul><ul><li>- Converting to decimal number. </li></ul>
  22. 22. Proposed Solution: Interface Design <ul><li>- Main page: </li></ul>
  23. 23. Proposed Solution: Interface Design(2) - Example of images Database:
  24. 24. Proposed Solution: Interface Design(3) <ul><li>- Scanning Process: </li></ul>
  25. 25. Proposed Solution: Interface Design(4) <ul><li>- Matched Image: </li></ul>
  26. 26. Proposed Solution: Interface Design(5) <ul><li>- Image with no matches: </li></ul>
  27. 27. Proposed Solution: Implementation Plan (Phase II) <ul><li>- Getting more information about MatLab. </li></ul><ul><li>- Studying tutorials about MatLab. </li></ul><ul><li>- Improving the Interim report details. </li></ul><ul><li>- Training and testing the system. </li></ul><ul><li>- Overcoming the problems. </li></ul><ul><li>- Understanding and analyzing the idea of Local Binary Patterns. </li></ul><ul><li>- More Research about Face Recognition based on Local Features. </li></ul>
  28. 28. Conclusion <ul><li>- LBP (Local Binary Patterns) is used to extract the local features in the face and match it with the most similar face image in the database. </li></ul><ul><li>- LBP is a method that works by dividing the face image to several blocks. </li></ul><ul><li>- In the matrix we compare the pixels with the center pixel . </li></ul><ul><li>- at the end we will get a binary number which will be converted into decimal format. </li></ul><ul><li>- will be combined together under one vector which will help to recognize the face. </li></ul>
  29. 29. <ul><li>The End </li></ul>

×