2. Sebastian Thrun CS223B Computer Vision, Winter 2005 1 Stanford CS223B Computer Vision, Winter 2006 Lecture 3 More On Features Professor Sebastian Thrun CAs: Dan Maynes-Aminzade and Mitul Saha [with slides by D Forsyth, D. Lowe, M. Polleyfeys, C. Rasmussen, G. Loy, D. Jacobs, J. Rehg, A, Hanson, G. Bradski,…]

3. Sebastian Thrun CS223B Computer Vision, Winter 2005 2 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features

4. Sebastian Thrun CS223B Computer Vision, Winter 2005 3 Features in Matlab im = imread('bridge.jpg'); bw = double(im(:,:,1))./ 256; edge(im,’sobel’) - (almost) linear edge(im,’canny’) - not local, no closed form

5. Sebastian Thrun CS223B Computer Vision, Winter 2005 4 Sobel Operator -1 -2 -1 0 0 0 1 2 1 -1 0 1 -2 0 2 -1 0 1 S1=S1=S 2 = Edge Magnitude = Edge Direction = S 1 + S 1 2 2 tan -1 S1S1 S2S2

6. Sebastian Thrun CS223B Computer Vision, Winter 2005 5 Sobel in Matlab edge(im,’sobel’)

7. Sebastian Thrun CS223B Computer Vision, Winter 2005 6 Canny Edge Detector edge(im,’canny’)

8. Sebastian Thrun CS223B Computer Vision, Winter 2005 7 Comparison CannySobel

9. Sebastian Thrun CS223B Computer Vision, Winter 2005 8 Canny Edge Detection Steps: 1.Apply derivative of Gaussian 2.Non-maximum suppression Thin multi-pixel wide “ridges” down to single pixel width 3.Linking and thresholding Low, high edge-strength thresholds Accept all edges over low threshold that are connected to edge over high threshold

10. Sebastian Thrun CS223B Computer Vision, Winter 2005 9 Non-Maximum Supression Non-maximum suppression: Select the single maximum point across the width of an edge.

11. Sebastian Thrun CS223B Computer Vision, Winter 2005 10 Linking to the Next Edge Point Assume the marked point q is an edge point. Take the normal to the gradient at that point and use this to predict continuation points (either r or p).

12. Sebastian Thrun CS223B Computer Vision, Winter 2005 11 Edge Hysteresis Hysteresis: A lag or momentum factor Idea: Maintain two thresholds k high and k low –Use k high to find strong edges to start edge chain –Use k low to find weak edges which continue edge chain Typical ratio of thresholds is roughly k high / k low = 2

13. Sebastian Thrun CS223B Computer Vision, Winter 2005 12 Canny Edge Detection (Example) courtesy of G. Loy gap is gone Original image Strong edges only Strong + connected weak edges Weak edges

14. Sebastian Thrun CS223B Computer Vision, Winter 2005 13 Canny Edge Detection (Example) Using Matlab with default thresholds

15. Sebastian Thrun CS223B Computer Vision, Winter 2005 14 Bridge Example Again edge(im,’canny’)

16. Sebastian Thrun CS223B Computer Vision, Winter 2005 15 Corner Effects

17. Sebastian Thrun CS223B Computer Vision, Winter 2005 16 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features

18. Sebastian Thrun CS223B Computer Vision, Winter 2005 17 Finding Corners Edge detectors perform poorly at corners. Corners provide repeatable points for matching, so are worth detecting. Idea: Exactly at a corner, gradient is ill defined. However, in the region around a corner, gradient has two or more different values.

19. Sebastian Thrun CS223B Computer Vision, Winter 2005 18 The Harris corner detector Form the second-moment matrix: Sum over a small region around the hypothetical corner Gradient with respect to x, times gradient with respect to y Matrix is symmetric Slide credit: David Jacobs

20. Sebastian Thrun CS223B Computer Vision, Winter 2005 19 Simple Case First, consider case where: This means dominant gradient directions align with x or y axis If either λ is close to 0, then this is not a corner, so look for locations where both are large. Slide credit: David Jacobs

21. Sebastian Thrun CS223B Computer Vision, Winter 2005 20 General Case It can be shown that since C is rotationally symmetric: So every case is like a rotated version of the one on last slide. Slide credit: David Jacobs

22. Sebastian Thrun CS223B Computer Vision, Winter 2005 21 So, To Detect Corners Filter image with Gaussian to reduce noise Compute magnitude of the x and y gradients at each pixel Construct C in a window around each pixel (Harris uses a Gaussian window – just blur) Solve for product of s (determinant of C) If s are both big (product reaches local maximum and is above threshold), we have a corner (Harris also checks that ratio of s is not too high)

23. Sebastian Thrun CS223B Computer Vision, Winter 2005 22 Gradient Orientation Closeup

24. Sebastian Thrun CS223B Computer Vision, Winter 2005 23 Corner Detection Corners are detected where the product of the ellipse axis lengths reaches a local maximum.

25. Sebastian Thrun CS223B Computer Vision, Winter 2005 24 Harris Corners Originally developed as features for motion tracking Greatly reduces amount of computation compared to tracking every pixel Translation and rotation invariant (but not scale invariant)

26. Sebastian Thrun CS223B Computer Vision, Winter 2005 25 Harris Corner in Matlab % Harris Corner detector - by Kashif Shahzad sigma=2; thresh=0.1; sze=11; disp=0; % Derivative masks dy = [-1 0 1; -1 0 1; -1 0 1]; dx = dy'; %dx is the transpose matrix of dy % Ix and Iy are the horizontal and vertical edges of image Ix = conv2(bw, dx, 'same'); Iy = conv2(bw, dy, 'same'); % Calculating the gradient of the image Ix and Iy g = fspecial('gaussian',max(1,fix(6*sigma)), sigma); Ix2 = conv2(Ix.^2, g, 'same'); % Smoothed squared image derivatives Iy2 = conv2(Iy.^2, g, 'same'); Ixy = conv2(Ix.*Iy, g, 'same'); % My preferred measure according to research paper cornerness = (Ix2.*Iy2 - Ixy.^2)./(Ix2 + Iy2 + eps); % We should perform nonmaximal suppression and threshold mx = ordfilt2(cornerness,sze^2,ones(sze)); % Grey-scale dilate cornerness = (cornerness==mx)&(cornerness>thresh); % Find maxima [rws,cols] = find(cornerness); % Find row,col coords. clf ; imshow(bw); hold on; p=[cols rws]; plot(p(:,1),p(:,2),'or'); title('\bf Harris Corners')

27. Sebastian Thrun CS223B Computer Vision, Winter 2005 26 Example (  =0.1)

28. Sebastian Thrun CS223B Computer Vision, Winter 2005 27 Example (  =0.01)

29. Sebastian Thrun CS223B Computer Vision, Winter 2005 28 Example (  =0.001)

30. Sebastian Thrun CS223B Computer Vision, Winter 2005 29 Harris: OpenCV Implementation

31. Sebastian Thrun CS223B Computer Vision, Winter 2005 30 Harris Corners for Optical Flow Will talk about this later (optical flow)

32. Sebastian Thrun CS223B Computer Vision, Winter 2005 31 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features

33. Sebastian Thrun CS223B Computer Vision, Winter 2005 32 Features? Local versus global

34. Sebastian Thrun CS223B Computer Vision, Winter 2005 33 Vanishing Points

35. Sebastian Thrun CS223B Computer Vision, Winter 2005 34 Vanishing Points

36. Sebastian Thrun CS223B Computer Vision, Winter 2005 35 Vanishing Points A. Canaletto [1740], Arrival of the French Ambassador in Venice

37. Sebastian Thrun CS223B Computer Vision, Winter 2005 36 Vanishing Points…? A. Canaletto [1740], Arrival of the French Ambassador in Venice

38. Sebastian Thrun CS223B Computer Vision, Winter 2005 37 From Edges to Lines

39. Sebastian Thrun CS223B Computer Vision, Winter 2005 38 Hough Transform

40. Sebastian Thrun CS223B Computer Vision, Winter 2005 39 Hough Transform: Quantization m m Detecting Lines by finding maxima / clustering in parameter space

41. Sebastian Thrun CS223B Computer Vision, Winter 2005 40 Hough Transform: Algorithm For each image point, determine –most likely line parameters b,m (direction of gradient) –strength (magnitude of gradient) Increment parameter counter by strength value Cluster in parameter space, pick local maxima

42. Sebastian Thrun CS223B Computer Vision, Winter 2005 41 Hough Transform: Results Hough TransformImageEdge detection

43. Sebastian Thrun CS223B Computer Vision, Winter 2005 42 Hough Transform?

44. Sebastian Thrun CS223B Computer Vision, Winter 2005 43 Summary Hough Transform Smart counting –Local evidence for global features –Organized in a table –Careful with parameterization! Subject to Curse of Dimensionality –Works great for simple features with 3 unknowns –Will fail for complex objects (e.g., all faces)

45. Sebastian Thrun CS223B Computer Vision, Winter 2005 44 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features

46. Sebastian Thrun CS223B Computer Vision, Winter 2005 45 Problem: Features for Recognition Want to find … in here

47. Sebastian Thrun CS223B Computer Vision, Winter 2005 46 Templates Find an object in an image! Want Invariance! –Scaling –Rotation –Illumination –Deformation

48. Sebastian Thrun CS223B Computer Vision, Winter 2005 47 Convolution with Templates % read image im = imread('bridge.jpg'); bw = double(im(:,:,1))./ 256;; imshow(bw) % apply FFT FFTim = fft2(bw); bw2 = real(ifft2(FFTim)); imshow(bw2) % define a kernel kernel=zeros(size(bw)); kernel(1, 1) = 1; kernel(1, 2) = -1; FFTkernel = fft2(kernel); % apply the kernel and check out the result FFTresult = FFTim.* FFTkernel; result = real(ifft2(FFTresult)); imshow(result) % select an image patch patch = bw(221:240,351:370); imshow(patch) patch = patch - (sum(sum(patch)) / size(patch,1) / size(patch, 2)); kernel=zeros(size(bw)); kernel(1:size(patch,1),1:size(patch,2)) = patch; FFTkernel = fft2(kernel); % apply the kernel and check out the result FFTresult = FFTim.* FFTkernel; result = max(0, real(ifft2(FFTresult))); result = result./ max(max(result)); result = (result.^ 1 > 0.5); imshow(result) % alternative convolution imshow(conv2(bw, patch, 'same'))

49. Sebastian Thrun CS223B Computer Vision, Winter 2005 48 Template Convolution

50. Sebastian Thrun CS223B Computer Vision, Winter 2005 49 Template Convolution

51. Sebastian Thrun CS223B Computer Vision, Winter 2005 50 Aside: Convolution Theorem Fourier Transform of g : F is invertible

52. Sebastian Thrun CS223B Computer Vision, Winter 2005 51 Convolution with Templates % read image im = imread('bridge.jpg'); bw = double(im(:,:,1))./ 256;; imshow(bw) % apply FFT FFTim = fft2(bw); bw2 = real(ifft2(FFTim)); imshow(bw2) % define a kernel kernel=zeros(size(bw)); kernel(1, 1) = 1; kernel(1, 2) = -1; FFTkernel = fft2(kernel); % apply the kernel and check out the result FFTresult = FFTim.* FFTkernel; result = real(ifft2(FFTresult)); imshow(result) % select an image patch patch = bw(221:240,351:370); imshow(patch) patch = patch - (sum(sum(patch)) / size(patch,1) / size(patch, 2)); kernel=zeros(size(bw)); kernel(1:size(patch,1),1:size(patch,2)) = patch; FFTkernel = fft2(kernel); % apply the kernel and check out the result FFTresult = FFTim.* FFTkernel; result = max(0, real(ifft2(FFTresult))); result = result./ max(max(result)); result = (result.^ 1 > 0.5); imshow(result) % alternative convolution imshow(conv2(bw, patch, 'same'))

53. Sebastian Thrun CS223B Computer Vision, Winter 2005 52 Convolution with Templates Invariances: –Scaling –Rotation –Illumination –Deformation Provides –Good localization No Maybe No

54. Sebastian Thrun CS223B Computer Vision, Winter 2005 53 Scale Invariance: Image Pyramid

55. Sebastian Thrun CS223B Computer Vision, Winter 2005 54 Aliasing

56. Sebastian Thrun CS223B Computer Vision, Winter 2005 55 Aliasing Effects Constructing a pyramid by taking every second pixel leads to layers that badly misrepresent the top layer

57. Sebastian Thrun CS223B Computer Vision, Winter 2005 56 Solution to Aliasing Convolve with Gaussian

58. Sebastian Thrun CS223B Computer Vision, Winter 2005 57 Templates with Image Pyramid Invariances: –Scaling –Rotation –Illumination –Deformation Provides –Good localization No Yes No Maybe No

59. Sebastian Thrun CS223B Computer Vision, Winter 2005 58 Template Matching, Commercial http://www.seeingmachines.com/facelab.htm

60. Sebastian Thrun CS223B Computer Vision, Winter 2005 59 Templates ?

61. Sebastian Thrun CS223B Computer Vision, Winter 2005 60 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features

62. Sebastian Thrun CS223B Computer Vision, Winter 2005 61 Let’s Return to this Problem… Want to find … in here

63. Sebastian Thrun CS223B Computer Vision, Winter 2005 62 SIFT Invariances: –Scaling –Rotation –Illumination –Deformation Provides –Good localization Yes Maybe Yes

64. Sebastian Thrun CS223B Computer Vision, Winter 2005 63 SIFT Reference Distinctive image features from scale-invariant keypoints. David G. Lowe, International Journal of Computer Vision, 60, 2 (2004), pp. 91-110. SIFT = Scale Invariant Feature Transform

65. Sebastian Thrun CS223B Computer Vision, Winter 2005 64 Invariant Local Features Image content is transformed into local feature coordinates that are invariant to translation, rotation, scale, and other imaging parameters SIFT Features

66. Advantages of invariant local features Locality: features are local, so robust to occlusion and clutter (no prior segmentation) Distinctiveness: individual features can be matched to a large database of objects Quantity: many features can be generated for even small objects Efficiency: close to real-time performance Extensibility: can easily be extended to wide range of differing feature types, with each adding robustness

67. Sebastian Thrun CS223B Computer Vision, Winter 2005 66 SIFT On-A-Slide 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates 2.Localizable corner: For each maximum fit quadratic function. Compute center with sub-pixel accuracy by setting first derivative to zero. 3.Eliminate edges: Compute ratio of eigenvalues, drop keypoints for which this ratio is larger than a threshold. 4.Enforce invariance to orientation: Compute orientation, to achieve scale invariance, by finding the strongest second derivative direction in the smoothed image (possibly multiple orientations). Rotate patch so that orientation points up. 5.Compute feature signature: Compute a "gradient histogram" of the local image region in a 4x4 pixel region. Do this for 4x4 regions of that size. Orient so that largest gradient points up (possibly multiple solutions). Result: feature vector with 128 values (15 fields, 8 gradients). 6.Enforce invariance to illumination change and camera saturation: Normalize to unit length to increase invariance to illumination. Then threshold all gradients, to become invariant to camera saturation.

68. Sebastian Thrun CS223B Computer Vision, Winter 2005 67 SIFT On-A-Slide 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates 2.Localizable corner: For each maximum fit quadratic function. Compute center with sub-pixel accuracy by setting first derivative to zero. 3.Eliminate edges: Compute ratio of eigenvalues, drop keypoints for which this ratio is larger than a threshold. 4.Enforce invariance to orientation: Compute orientation, to achieve scale invariance, by finding the strongest second derivative direction in the smoothed image (possibly multiple orientations). Rotate patch so that orientation points up. 5.Compute feature signature: Compute a "gradient histogram" of the local image region in a 4x4 pixel region. Do this for 4x4 regions of that size. Orient so that largest gradient points up (possibly multiple solutions). Result: feature vector with 128 values (15 fields, 8 gradients). 6.Enforce invariance to illumination change and camera saturation: Normalize to unit length to increase invariance to illumination. Then threshold all gradients, to become invariant to camera saturation.

69. Sebastian Thrun CS223B Computer Vision, Winter 2005 68 Find Invariant Corners 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates

70. Sebastian Thrun CS223B Computer Vision, Winter 2005 69 Finding “Keypoints” (Corners) Idea: Find Corners, but scale invariance Approach: Run linear filter (diff of Gaussians) At different resolutions of image pyramid

71. Sebastian Thrun CS223B Computer Vision, Winter 2005 70 Difference of Gaussians Minus Equals

72. Sebastian Thrun CS223B Computer Vision, Winter 2005 71 Difference of Gaussians surf(fspecial('gaussian',40,4)) surf(fspecial('gaussian',40,8)) surf(fspecial('gaussian',40,8) - fspecial('gaussian',40,4))

73. Sebastian Thrun CS223B Computer Vision, Winter 2005 72 Find Corners with DiffOfGauss im =imread('bridge.jpg'); bw = double(im(:,:,1)) / 256; for i = 1 : 10 gaussD = fspecial('gaussian',40,2*i) - fspecial('gaussian',40,i); res = abs(conv2(bw, gaussD, 'same')); res = res / max(max(res)); imshow(res) ; title(['\bf i = ' num2str(i)]); drawnow end

74. Sebastian Thrun CS223B Computer Vision, Winter 2005 73 Gaussian Kernel Size i=1

75. Sebastian Thrun CS223B Computer Vision, Winter 2005 74 Gaussian Kernel Size i=2

76. Sebastian Thrun CS223B Computer Vision, Winter 2005 75 Gaussian Kernel Size i=3

77. Sebastian Thrun CS223B Computer Vision, Winter 2005 76 Gaussian Kernel Size i=4

78. Sebastian Thrun CS223B Computer Vision, Winter 2005 77 Gaussian Kernel Size i=5

79. Sebastian Thrun CS223B Computer Vision, Winter 2005 78 Gaussian Kernel Size i=6

80. Sebastian Thrun CS223B Computer Vision, Winter 2005 79 Gaussian Kernel Size i=7

81. Sebastian Thrun CS223B Computer Vision, Winter 2005 80 Gaussian Kernel Size i=8

82. Sebastian Thrun CS223B Computer Vision, Winter 2005 81 Gaussian Kernel Size i=9

83. Sebastian Thrun CS223B Computer Vision, Winter 2005 82 Gaussian Kernel Size i=10

84. Sebastian Thrun CS223B Computer Vision, Winter 2005 83 Key point localization Detect maxima and minima of difference-of-Gaussian in scale space

85. Sebastian Thrun CS223B Computer Vision, Winter 2005 84 Example of keypoint detection (a) 233x189 image (b) 832 DOG extrema (c) 729 above threshold

86. Sebastian Thrun CS223B Computer Vision, Winter 2005 85 SIFT On-A-Slide 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates 2.Localizable corner: For each maximum fit quadratic function. Compute center with sub-pixel accuracy by setting first derivative to zero. 3.Eliminate edges: Compute ratio of eigenvalues, drop keypoints for which this ratio is larger than a threshold. 4.Enforce invariance to orientation: Compute orientation, to achieve scale invariance, by finding the strongest second derivative direction in the smoothed image (possibly multiple orientations). Rotate patch so that orientation points up. 5.Compute feature signature: Compute a "gradient histogram" of the local image region in a 4x4 pixel region. Do this for 4x4 regions of that size. Orient so that largest gradient points up (possibly multiple solutions). Result: feature vector with 128 values (15 fields, 8 gradients). 6.Enforce invariance to illumination change and camera saturation: Normalize to unit length to increase invariance to illumination. Then threshold all gradients, to become invariant to camera saturation.

87. Sebastian Thrun CS223B Computer Vision, Winter 2005 86 Example of keypoint detection Threshold on value at DOG peak and on ratio of principle curvatures (Harris approach) (c) 729 left after peak value threshold (from 832) (d) 536 left after testing ratio of principle curvatures

88. Sebastian Thrun CS223B Computer Vision, Winter 2005 87 SIFT On-A-Slide 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates 2.Localizable corner: For each maximum fit quadratic function. Compute center with sub-pixel accuracy by setting first derivative to zero. 3.Eliminate edges: Compute ratio of eigenvalues, drop keypoints for which this ratio is larger than a threshold. 4.Enforce invariance to orientation: Compute orientation, to achieve scale invariance, by finding the strongest second derivative direction in the smoothed image (possibly multiple orientations). Rotate patch so that orientation points up. 5.Compute feature signature: Compute a "gradient histogram" of the local image region in a 4x4 pixel region. Do this for 4x4 regions of that size. Orient so that largest gradient points up (possibly multiple solutions). Result: feature vector with 128 values (15 fields, 8 gradients). 6.Enforce invariance to illumination change and camera saturation: Normalize to unit length to increase invariance to illumination. Then threshold all gradients, to become invariant to camera saturation.

89. Sebastian Thrun CS223B Computer Vision, Winter 2005 88 Select canonical orientation Create histogram of local gradient directions computed at selected scale Assign canonical orientation at peak of smoothed histogram Each key specifies stable 2D coordinates (x, y, scale, orientation)

90. Sebastian Thrun CS223B Computer Vision, Winter 2005 89 SIFT On-A-Slide 1.Enforce invariance to scale: Compute Gaussian difference max, for may different scales; non-maximum suppression, find local maxima: keypoint candidates 2.Localizable corner: For each maximum fit quadratic function. Compute center with sub-pixel accuracy by setting first derivative to zero. 3.Eliminate edges: Compute ratio of eigenvalues, drop keypoints for which this ratio is larger than a threshold. 4.Enforce invariance to orientation: Compute orientation, to achieve scale invariance, by finding the strongest second derivative direction in the smoothed image (possibly multiple orientations). Rotate patch so that orientation points up. 5.Compute feature signature: Compute a "gradient histogram" of the local image region in a 4x4 pixel region. Do this for 4x4 regions of that size. Orient so that largest gradient points up (possibly multiple solutions). Result: feature vector with 128 values (15 fields, 8 gradients). 6.Enforce invariance to illumination change and camera saturation: Normalize to unit length to increase invariance to illumination. Then threshold all gradients, to become invariant to camera saturation.

91. Sebastian Thrun CS223B Computer Vision, Winter 2005 90 SIFT vector formation Thresholded image gradients are sampled over 16x16 array of locations in scale space Create array of orientation histograms 8 orientations x 4x4 histogram array = 128 dimensions

92. Sebastian Thrun CS223B Computer Vision, Winter 2005 91 Nearest-neighbor matching to feature database Hypotheses are generated by approximate nearest neighbor matching of each feature to vectors in the database –SIFT use best-bin-first (Beis & Lowe, 97) modification to k-d tree algorithm –Use heap data structure to identify bins in order by their distance from query point Result: Can give speedup by factor of 1000 while finding nearest neighbor (of interest) 95% of the time

93. Sebastian Thrun CS223B Computer Vision, Winter 2005 92 3D Object Recognition Extract outlines with background subtraction

94. Sebastian Thrun CS223B Computer Vision, Winter 2005 93 3D Object Recognition Only 3 keys are needed for recognition, so extra keys provide robustness Affine model is no longer as accurate

95. Sebastian Thrun CS223B Computer Vision, Winter 2005 94 Recognition under occlusion

96. Sebastian Thrun CS223B Computer Vision, Winter 2005 95 Test of illumination invariance Same image under differing illumination 273 keys verified in final match

97. Sebastian Thrun CS223B Computer Vision, Winter 2005 96 Examples of view interpolation

98. Sebastian Thrun CS223B Computer Vision, Winter 2005 97 Location recognition

99. Sebastian Thrun CS223B Computer Vision, Winter 2005 98 SIFT Invariances: –Scaling –Rotation –Illumination –Deformation Provides –Good localization Yes Maybe Yes

100. Sebastian Thrun CS223B Computer Vision, Winter 2005 99 Today’s Goals Canny Edge Detector Harris Corner Detector Hough Transform Templates and Image Pyramid SIFT Features