1. Miniature faking http://en.wikipedia.org/wiki/File:Jodhpur_tilt_shift.jpg In close-up photo, the depth of field is limited.

2. Miniature faking

3. http://en.wikipedia.org/wiki/File:Oregon_State_Beavers_Tilt-Shift_Miniature_Greg_Keene.jpg

4. Review Previous section: – Feature detection and matching – Model fitting and outlier rejection

5. Project 2 discussion after break

6. Review: Interest points Keypoint detection: repeatable and distinctive – Corners, blobs, stable regions – Harris, DoG, MSER

7. Harris Detector [ Harris88 ] Second moment matrix 7 1. Image derivatives 2. Square of derivatives 3. Gaussian filter g(  I ) IxIx IyIy Ix2Ix2 Iy2Iy2 IxIyIxIy g(I x 2 )g(I y 2 ) g(I x I y ) 4. Cornerness function – both eigenvalues are strong har 5. Non-maxima suppression (optionally, blur first)

8. Review: Local Descriptors Most features can be thought of as templates, histograms (counts), or combinations Most available descriptors focus on edge/gradient information – Capture texture information – Color rarely used K. Grauman, B. Leibe

9. x y b m x ym 353322 37111043 231452 210133 b Slide from S. Savarese Review: Hough transform

10. Review: RANSAC Algorithm: 1. Sample (randomly) the number of points required to fit the model (#=2) 2. Solve for model parameters using samples 3. Score by the fraction of inliers within a preset threshold of the model Repeat 1-3 until the best model is found with high confidence

11. Review: 2D image transformations Szeliski 2.1

12. This section – multiple views Today – Intro and stereo Wednesday – Camera calibration Monday – Fundamental Matrix

13. Computer Vision James Hays Stereo: Intro Slides by Kristen Grauman

14. Multiple views Hartley and Zisserman Lowe stereo vision structure from motion optical flow

15. Why multiple views? Structure and depth are inherently ambiguous from single views. Images from Lana Lazebnik

16. Why multiple views? Structure and depth are inherently ambiguous from single views. Optical center P1 P2 P1’=P2’

17. What cues help us to perceive 3d shape and depth?

18. Shading [Figure from Prados & Faugeras 2006]

19. Focus/defocus [figs from H. Jin and P. Favaro, 2002] Images from same point of view, different camera parameters 3d shape / depth estimates

20. Texture [From A.M. Loh. The recovery of 3-D structure using visual texture patterns. PhD thesis]A.M. Loh. The recovery of 3-D structure using visual texture patterns.

21. Perspective effects Image credit: S. Seitz

22. Motion Figures from L. Zhanghttp://www.brainconnection.com/teasers/?main=illusion/motion-shape

23. Occlusion Rene Magritt'e famous painting Le Blanc-Seing (literal translation: "The Blank Signature") roughly translates as "free hand" or "free rein".

24. If stereo were critical for depth perception, navigation, recognition, etc., then this would be a problem

25. Multi-view geometry problems Structure: Given projections of the same 3D point in two or more images, compute the 3D coordinates of that point Camera 3 R 3,t 3 Slide credit: Noah Snavely ? Camera 1 Camera 2 R 1,t 1 R 2,t 2

26. Multi-view geometry problems Stereo correspondence: Given a point in one of the images, where could its corresponding points be in the other images? Camera 3 R 3,t 3 Camera 1 Camera 2 R 1,t 1 R 2,t 2 Slide credit: Noah Snavely

27. Multi-view geometry problems Motion: Given a set of corresponding points in two or more images, compute the camera parameters Camera 1 Camera 2 Camera 3 R 1,t 1 R 2,t 2 R 3,t 3 ? ? ? Slide credit: Noah Snavely

28. Estimating scene shape “Shape from X”: Shading, Texture, Focus, Motion… Stereo: –shape from “motion” between two views –infer 3d shape of scene from two (multiple) images from different viewpoints scene point optical center image plane Main idea:

29. Human eye Fig from Shapiro and Stockman Pupil/Iris – control amount of light passing through lens Retina - contains sensor cells, where image is formed Fovea – highest concentration of cones Rough analogy with human visual system:

30. Human stereopsis: disparity Human eyes fixate on point in space – rotate so that corresponding images form in centers of fovea.

31. Disparity occurs when eyes fixate on one object; others appear at different visual angles Human stereopsis: disparity

32. Disparity: d = r-l = D-F. Human stereopsis: disparity Forsyth & Ponce

33. Random dot stereograms Julesz 1960: Do we identify local brightness patterns before fusion (monocular process) or after (binocular)? To test: pair of synthetic images obtained by randomly spraying black dots on white objects

34. Random dot stereograms Forsyth & Ponce

35. Random dot stereograms

36. When viewed monocularly, they appear random; when viewed stereoscopically, see 3d structure. Human binocular fusion not directly associated with the physical retinas; must involve the central nervous system (V2, for instance). Imaginary “cyclopean retina” that combines the left and right image stimuli as a single unit High level scene understanding not required for Stereo But, high level scene understanding is arguably better than stereo

37. Stereo photography and stereo viewers Invented by Sir Charles Wheatstone, 1838 Image from fisher-price.com Take two pictures of the same subject from two slightly different viewpoints and display so that each eye sees only one of the images.

38. http://www.johnsonshawmuseum.org

40. Public Library, Stereoscopic Looking Room, Chicago, by Phillips, 1923

41. http://www.well.com/~jimg/stereo/stereo_list.html

42. Autostereograms Images from magiceye.com Exploit disparity as depth cue using single image. (Single image random dot stereogram, Single image stereogram)

43. Images from magiceye.com Autostereograms

44. Estimating depth with stereo Stereo: shape from “motion” between two views We’ll need to consider: Info on camera pose (“calibration”) Image point correspondences scene point optical center image plane

45. Two cameras, simultaneous views Single moving camera and static scene Stereo vision

46. Camera parameters Camera frame 1 Intrinsic parameters: Image coordinates relative to camera  Pixel coordinates Extrinsic parameters: Camera frame 1  Camera frame 2 Camera frame 2 Extrinsic params: rotation matrix and translation vector Intrinsic params: focal length, pixel sizes (mm), image center point, radial distortion parameters We’ll assume for now that these parameters are given and fixed.

47. Geometry for a simple stereo system First, assuming parallel optical axes, known camera parameters (i.e., calibrated cameras):

48. baseline optical center (left) optical center (right) Focal length World point image point (left) image point (right) Depth of p

49. Assume parallel optical axes, known camera parameters (i.e., calibrated cameras). What is expression for Z? Similar triangles (p l, P, p r ) and (O l, P, O r ): Geometry for a simple stereo system disparity

50. Depth from disparity image I(x,y) image I´(x´,y´) Disparity map D(x,y) (x´,y´)=(x+D(x,y), y) So if we could find the corresponding points in two images, we could estimate relative depth…

51. Break Feel free to leave if you’ve got Project 2 mostly working

52. proj2 Common Problems Interest point detection Feature description Matching

53. Interest Point Detection Suppress corners on the edge of the image which might have appeared from padding artifacts You probably shouldn’t have too few (50) or too many (5000) interest points. Colfilt() approach to non-max suppression may be more reliable. –But I’ve seen people using it incorrectly.

54. Feature Description You can start with something very simple, e.g. cutting out raw image patches. –My implementation gets 74/100 matches correct with this, versus 93/100 with a more SIFT-like feature. You can use imagesc(image1_features) to check that the features aren’t degenerate in some way. Free parameters (e.g. size of Gaussians) can matter