Ders1- Introduction to Computer Vision.ppt

Introduction to Computer Vision

1. Computer Vision
Prof. Songül Varlı
Yildiz Technical University
Computer Engineering Department
Course Contents
Introduction to Computer Vision

2. Computer Vision
What we see
What a computer sees

3. Lighting
Scene
Camera
Computer
Scene Interpretation
Components of a Computer Vision System

4. Sampling and Quantization

5. What is a digital image?
We usually operate on digital (discrete)images:
Samplethe 2D space on a regular grid
Quantizeeach sample (round to nearest integer)
If our samples are Dapart, we can write this as:
f[i,j] = Quantize{ f(iD, jD) }
The image can now be represented as a matrix of integer values

6. Steps of Image Processing

7. Negative Image

8. Image Enhancement

9. Intensity Transformation
Gamma Correction

10. Contrast Enhancement

11. Image Histogram

12. Application Areas of
Image Processing

13. Document Handling
[Intelligent Document Recognition]

14. Biometrics

15. Fingerprint Identification/Authentication

16. Interpretation of AerialPhotography

17. Traffic Monitoring

18. Tracking

19. Microsoft Kinect
IR Camera
RGB Camera
IR LED Emitter

20. Face Detection, Viola & Jones,2001

21. Face Recognition
•Principle Components Analysis (PCA)
•Face Recognition

22. Facial Expression Recognition

23. Hand Posture/Gesture Recognition
Smart Human-Computer User Interfaces
Sign Language Recognition

24. Human Activity Recognition

25. Medical Applications
skin cancer breast cancer


26. SIFT & Object Recognition,
David Lowe, 1999

27. Histogram of Oriented Gradients
(HoG) Dalal & Triggs, 2005

28. PASCAL Visual Object Challenge
(20 object categories)
[Everingham et al. 2006-2012]

29. Segmentation accuracy by class over
the yearsfor PASCAL Visual Object
Challenge

30. The Image Classification Challenge:
1,000 object classes
1,431,167 images
Russakovsky et al. arXiv, 2014

31. The Image Classification Challenge:
1,000 object classes
1,431,167 images
Error Rates are given in graph

32. Object detection & Image captioning

33. Convolutional Neural Networks
(CNN)

34. Convolutional Neural Networks (CNN)
were not invented overnight