Distinctive Image Features from ScaleInvariant Keypoints Ronnie Bajwa

Distinctive Image Features from Scale-Invariant Keypoints Ronnie Bajwa Sameer Pawar * * Adapted from slides found online by Michael Kowalski, Lehigh University

Goal l Extract distinctive invariant features from an image that can be used to perform reliable object recognition and many other applications.

Key requirements for good feature l l l Highly distinctive, with a low probability of mismatch Should be easy to extract Invariance: - Scale and rotation - change in illumination - slight change in viewing angle - image noise - clutter and occlusion

Brief History l Moravec(1981): – l Harris(1988): – – l Selects locations that has large gradients in all directions (corners) Invariant to image rotation and to slight affine intensity change, but faces problems for scale change. Zhang(1995): – l corner detector. Introduced correlation window to match images in large range motion. Schmid (1997): – – – Suggested use of local feature matching for image recognition. Used rotationally invariant descriptor of the local image region. Multiple feature matches accomplish recognition under occlusion & clutter.

Harris Corner Detector: direction of the fastest change /2 )-1 ax ( m “Flat” 2 ~ 1~0 ( m “Edge” 2 >> 1 )-1 in /2 direction of the slowest change “Corner” 2 ~ 1 ~ Large

SIFT Algorithm Overview Filtered approach 1. Scale-space extrema detection – – 2. Keypoint localization – – 3. 4. Identify potential points: invariant to scale & orientation. Difference-of-Gaussian function Improve the estimate for location by fitting a quadratic Extrema thresholded for filter out insignificant points. Orientation Assignment Orientation assigned to each keypoint and neighboring pixels based on local gradient. Keypoint Descriptor construction – Feature vector based on gradients of local neighborhood

Keypoint Selection: Scale space l We express the image at different scales by filtering it with a Gaussian kernel

Keypoint Selection: why Do. G’s? l l l Lindeberg(1994) and Mikolajczyk (2002) found that the maxima and minima of the scaled Laplacian provides the most stable scale invariant features We can use the scaled images to approximate this: Efficient to compute – Smoothed images L needed later so D can be computed by simple image subtraction

Back to picking keypoints 1. 2. 3. 4. 5. Supersample original image Compute smoothed images using different scales σ for entire octave Compute do. G images from adjacent scales for entire octave Isolate keypoints in each octave by detecting extrema in do. G compared to neighboring pixels Subsample image 2σ of current octave and repeat process (2 -3) for next octave

Visual representation of process

Visual representation of process (cont. )

A More Visual Representation Original Image Starting Image

A More Visual Representation (cont. ) First Octave of L images

A More Visual Representation (cont. ) Second Octave Third Octave Fourth Octave

A More Visual Representation (cont. ) First Octave difference-of-Gaussians

A More Visual Representation (cont. )

Scale space sampling l How many fine scales in every octave? l Extremas can be arbitrary close but very close ones are unstable. After subsampling and before finding scaled images of the octave, prior smoothing of 1. 6 is done To compensate the loss of higher spatial frequencies, original image is doubled in size l l

Accurate Keypoint Localization l l l From difference-of-Gaussian local extrema detection we obtain approximate values for keypoints Originally these approximations were used directly For an improvement in matching and stability fitting to a 3 D quadratic function is used

The Taylor Series Expansion l Take Taylor Series Expansion of scale-space function D(x, y, σ) – Use up to quadratic terms – origin shifted to sample point offset from this sample point to find location of extremum, take derivative and set to 0 – –

Thresholding Keypoints (part 1) l The function value at the extrema is used to reject unstable extrema – – – Low contrast Evaluate Absolute value less than 0. 03 at extrema location results in discarding of extrema

Thresholding Keypoints (part 2) l Difference-of-Gaussian function will be strong along edges – – – Some locations along edges are poorly determined and will become unstable when even small amounts of noise are added These locations will have a large principal curvature across the edge but a small principal of curvature perpendicular to the edge Therefore we need to compute the principal curvatures at the location and compare the two

Computing the Principal Curvatures l Hessian matrix – The eigenvalues of H are proportional to principal curvatures – We are not concerned about actual values of eigenvalue, just the ratio of the two

Stages of keypoint selection

Assigning an Orientation l l We finally have a keypoint that we are going to keep The next step is assigning an orientation for the keypoint – Used in making the matching technique invariant to rotation

Assigning an Orientation (cont. ) l l l Gaussian smoothed image, L, with closest scale is chosen (scale invariance) Points in region around keypoint are selected and magnitude and orientations of gradient are calculated Orientation histogram formed with 36 bins. Sample is added to appropriate bin and weighted by gradient magnitude and Gaussian-weighted circular window with a of σ 1. 5 times scale of keypoint

Assigning an Orientation (cont. ) l l l Highest peak in orientation is found along with any other peaks within 80% of highest peak 3 closest histogram values to each peak are used to interpolate (fit to a parabola) a better accurate peak As of now each keypoint has 4 dimesions: x location, y location, scale, and orientation

Keypoint Descriptor l l Calculated using a region around the keypoint as opposed to directly from the keypoint for robustness Like before, magnitudes and orientations are calculated for points in region around keypoint using L of nearest scale To ensure orientation invariance the gradient orientations and coordinates of descriptor are rotated relative to orientation of keypoint Provides invariance to changes in illumination and 3 D camera viewpoint

Visual Representation of Keypoint Descriptor

Keypoint Descriptor (cont. ) l Magnitude of each point is weighted with σ of one half the width of the descriptor window – – l Stops sudden changes in descriptor due to small changes in position of the window Gives less weight to gradients far from keypoint Samples are divided into 4 x 4 subregions around keypoint. Allows for a change of up to 4 positions of a sample while still being included in the same histogram

Keypoint Descriptor (cont. ) l Avoiding boundary effects between histograms – Trilinear interpolation used to distribute value of gradient of each sample into adjacent histogram bins l l Weight equal to 1 – d , where d is the distance of a specific sample to the center of a bin Vector normalization – – – Done at the end to ensure invariance to illumination change (affine) Entire vector normalized to 1 To combat non-linear illumination (camera saturation) changes values in feature vector are thresholded to no larger than 0. 2 and then the vector is normalized.

We now have features l Up to this point we have: – – – l Found rough approximations for features by looking at the difference-of-Gaussians Localized the keypoint more accurately Removed poor keypoints Determined the orientation of a keypoint Calculated a 128 feature vector for each keypoint What do we do now?

Matching Keypoints between images l Tale of two images (or more) – – l l One image is the training sample of what we are looking for The other image is the world picture that might contain instances of the training sample Both images have features associated with them across different octaves How do we match features between the two?

Matching Keypoints between images (cont. ) l Nearest Neighbor Algorithm(Euclidean Distance) – – Independently match all keypoints in all octaves in one image with all keypoints in all octaves in other image How to solve problem of features that have no correct match with opposite image (i. e. background noise, clutter) l l Global threshold on distance (did not perform well) Ratio of closest neighbor with second closest neighbor.

Matching Keypoints between images (cont. ) l Threshold at 0. 8 – – Eliminates 90% of false matches Eliminates less than 5% of correct matches

Efficient Nearest Neighbor indexing l In high dimensions no algorithms – k-dimensional tree: Is O(2 klog n).

Efficient Nearest Neighbor indexing l Use an approximate algorithm called Best. Bin-First (feature space) – – Bins are searched in order of their closest distance. Heap-based priority queue. Returns closest neighbor with high probability Search cut off after searching the 200 nearest-neighbor candidates l Works well since we only consider matches where nearest neighbor is less than 0. 8 times the distance to second nearest neighbor.

Clustering -Hough Transform l Small Highly-occluded objects: – – – l 3 feature matches are sufficient. 99% outliers. RANSAC doesn’t perform well due to large number of outliers. Hough Transform – – – Uses each feature to vote for all object poses consistent with the feature Predicts approx. model for similarity transformation. Bin Sizes must be large to account for large error bounds l l l Orientation-30 degrees, Scale-of 2, Location-0. 25. Each keypoint votes to two closest bins in each dimension/ 16 entries for each hypothesis (feature).

Model Verification-Affine Transformation l Geometric Verification of clusters

Results (Object Recognition)

Results (cont. )

Questions?