540 likes | 1.74k Views
Image Pyramids. Computer Vision, winter 2012-13 CS Department, Technion. Topics. The Gaussian Pyramid The Laplacian Pyramid Applications: Pattern Matching Coding (Compression) Enhancement Blending. Gaussian Pyramid. Gaussian Pyramid. Approximation filter is not Gaussian
E N D
Image Pyramids Computer Vision, winter 2012-13 CS Department, Technion
Topics • The Gaussian Pyramid • The Laplacian Pyramid • Applications: • Pattern Matching • Coding (Compression) • Enhancement • Blending
Gaussian Pyramid • Approximation filter is not Gaussian • A 5 tap binomial kernel: • 2D extension:
Laplacian Pyramid • Approximation of the Laplacian of each scale • Constructed by Difference of Gaussian (DoG) layers
Step II Blend 1 Blend 1 Mask Blend 2 Blend 2 Mask • Weighted sum over each Laplacianpyramid scale (of the images) independently • Gaussian pyramid of masks, or simple resizing of masks to correct dimensions
Step III • Standard Laplacian pyramid reconstruction of blended scales
Implementation Issues • Image registration • Feathering • Over/under flows outside [0,1] • MATLAB function: impyramid() • Only Gaussian pyramid • ‘reduce’ or ‘expand’ modes • MATLAB function: imresize() • Useful for masks resizing