“Feature detection 2”

Image transformation

  1. Geometric: Rotation, Scale
  2. Photometric: Intensity change

Invariance and equivariance

We want corner locations to be invariant to photometric transformations and equivariant to geometric transformations

  1. Invariance: image is transformed and corner locations do not change
  2. Equavariance: if we have two transformed versions of the same image, features should be detected in corresponding locations

Some operartion and their characteristic

  1. Derivatives and window function are equivariant
  2. Corner location is equivariant w.r.t. translation
  3. Corner location is equivariant w.r.t. image rotation
  4. Affine intensity change: Partially invariant to affine intensity change
  5. Scaling is neither invariant nor equivariant to scaling

Key idea to scale invariant

find scale that gives local maximum of f

  1. in both position and scale
  2. one definition of f: the Harris operator
  3. Instead of computing f for larger and larger windows, we can implement using a fixed window size with a Gaussian pyramid

Laplacian of Gaussian

\(\nabla^{2} g=\frac{\partial^{2} g}{\partial x^{2}}+\frac{\partial^{2} g}{\partial y^{2}}\)

This is a “Blob” detector, the maxima and minima of LoG operator in space and scale are blobs in the image

An alternative

DoG: Difference of Gaussians

[D o G=G(x, y, k \sigma)-G(x, y, \sigma)]



Hi, I'm Yong Huang. I've recently graduated from Cornell Tech and obtained my master's degree, I shall start my Ph.D. in Computer Science this fall at UC Irvine. Thank you for visiting my site.