Scale Invariant Region Selection and SIFT Sung-Eui Yoon ( ) - - PowerPoint PPT Presentation

▶
scale invariant region selection and sift
SMART_READER_LITE
LIVE PREVIEW

Scale Invariant Region Selection and SIFT Sung-Eui Yoon ( ) - - PowerPoint PPT Presentation

Nov 13, 2022 585 likes •1.05k views

CS688: Web-Scale Image Search Scale Invariant Region Selection and SIFT Sung-Eui Yoon ( ) Course URL: http://sgvr.kaist.ac.kr/~sungeui/IR Announcements Parts of my book are updated One of students is invited to each class 5

slide-1
SLIDE 1

CS688: Web-Scale Image Search

Scale Invariant Region Selection and SIFT

Sung-Eui Yoon (윤성의)

Course URL: http://sgvr.kaist.ac.kr/~sungeui/IR

slide-2
SLIDE 2

5

Announcements

  • Parts of my book are updated
  • One of students is invited to each class
slide-3
SLIDE 3

6

Class Objectives (Ch. 2.4)

  • Scale invariant region selection
  • Automatic scale selection
  • Laplacian of Gradients (LoG)

Difference of Gradients (DoG)

  • SIFT as a local descriptor
  • At last time, we discussed:
  • Different conferences
  • Image descriptors that are invariant to various

changes

  • Harris corner detector
slide-4
SLIDE 4

7

slide-5
SLIDE 5

8

slide-6
SLIDE 6

9

slide-7
SLIDE 7

10

slide-8
SLIDE 8

11

slide-9
SLIDE 9

12

slide-10
SLIDE 10

13

slide-11
SLIDE 11

14

slide-12
SLIDE 12

15

slide-13
SLIDE 13

16

slide-14
SLIDE 14

17

slide-15
SLIDE 15

18

slide-16
SLIDE 16

19

slide-17
SLIDE 17

20

slide-18
SLIDE 18

21

slide-19
SLIDE 19

22

slide-20
SLIDE 20

23

slide-21
SLIDE 21

24

slide-22
SLIDE 22

25

slide-23
SLIDE 23

26

slide-24
SLIDE 24

27

slide-25
SLIDE 25

28

slide-26
SLIDE 26

29

slide-27
SLIDE 27

30

slide-28
SLIDE 28

31

slide-29
SLIDE 29

32

slide-30
SLIDE 30

33

slide-31
SLIDE 31

34

slide-32
SLIDE 32

35

slide-33
SLIDE 33

36

slide-34
SLIDE 34

37

slide-35
SLIDE 35

38

slide-36
SLIDE 36

39

slide-37
SLIDE 37

40

slide-38
SLIDE 38

41

slide-39
SLIDE 39

42

Other Descriptors

  • GIST: a kind of SIFT in a global scale
  • SURF: an acceleration using the integral

image, i.e., summed area table

  • CNN features
slide-40
SLIDE 40

43

80M Tiny Images

  • Just use 32 by 32 images
  • It works well even for recognition with a

simple recognition method (nearest neighbor search) with using 80M data

  • Indicates the importance of data
slide-41
SLIDE 41

44

PA1 (Optional)

  • Objective
  • Understand how to extract SIFT features and to

use related libraries (OpenCV, vlfeat, … )

slide-42
SLIDE 42

45

Class Objectives (Ch. 2.4) were:

  • Scale invariant region selection
  • Automatic scale selection
  • Laplacian of Gradients (LoG)

Difference of Gradients (DoG)

  • SIFT as a local descriptor
slide-43
SLIDE 43

46

Next Time…

  • Basic deep learning and its applications to

computer vision

  • Intro to object recognition
  • Bag-of-Words (BoW) models
slide-44
SLIDE 44

47

Homework for Every Class

  • Go over the next lecture slides
  • Come up with one question on what we have

discussed today

  • 1 for typical questions (that were answered in the class)
  • 2 for questions with thoughts or that surprised me
  • Write questions 3 times before the mid-term exam
  • Write a question about one out of every four classes
  • Multiple questions in one time will be counted as one time
  • Common questions are compiled at the Q&A file
  • Some of questions will be discussed in the class
  • If you want to know the answer of your question,

ask me or TA on person