Scribd
Upload
6 views2 pages

Image Piramides

Image_Piramides

Uploaded by

5bnvpv9db4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views2 pages

Image Piramides

Image_Piramides

Uploaded by

5bnvpv9db4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Image Pyramids

January 6, 2025

Image Pyramids
Image pyramids are hierarchical representations of images that capture information at multiple scales. They
are widely used in various image processing and computer vision tasks, offering benefits such as multi-
resolution analysis, feature extraction, and image compression.

Pyramid Representation
Image pyramids are structured as a series of layers or levels, with each level containing a version of the
original image at a different scale. The highest level of the pyramid typically represents the original image,
while subsequent levels capture progressively lower resolutions. This multi-scale representation allows for
efficient processing of images at different levels of detail.

Gaussian Image Pyramids


Gaussian image pyramids are constructed using a combination of Gaussian smoothing and downsampling
operations. The process involves the following steps:

1. Gaussian Smoothing: The original image is convolved with a Gaussian kernel to blur the image and
remove high-frequency noise.
2. Downsampling: The smoothed image is then downsampled to reduce its resolution. This is typically
done by selecting every other pixel in each row and column of the image.
3. Repeat: The process is repeated iteratively to create multiple levels of the pyramid, with each level
representing a progressively smaller version of the original image.

Properties
One important property of Gaussian pyramids is that they are used primarily for downsampling and image
compression. Due to the smoothing operation, information is lost during the downsampling process, making
it impossible to reconstruct the original image from the downsampled versions.

Laplacian Image Pyramid


Laplacian image pyramids are derived from Gaussian pyramids and are used for image reconstruction. Un-
like Gaussian pyramids, Laplacian pyramids allow for the reconstruction of the original image from the
downsampled versions. The construction process involves the following steps:

1. Gaussian Pyramid: First, a Gaussian pyramid is constructed as described earlier.


2. Upsampling: The downsampled images from the Gaussian pyramid are upsampled to restore their
original size.
3. Difference: The upsampled images are then subtracted from the original images to obtain the Lapla-
cian images.

1
Why is it called a Laplacian Pyramid?
The Laplacian images obtained from the subtraction process represent the Laplacian of the Gaussian-
smoothed images. This difference operation highlights the high-frequency details in the image, making
it possible to reconstruct the original image from the Laplacian pyramid.

Steerable Pyramid
A steerable pyramid is a multi-scale representation of an image that is designed to capture directional
information. Unlike Gaussian and Laplacian pyramids, which are based on isotropic (directionally invariant)
filters, steerable pyramids use steerable filters that can be rotated to different orientations. This allows
for the extraction of oriented features in images, making steerable pyramids useful for tasks such as edge
detection, texture analysis, and object recognition.

Wavelets
Wavelets are mathematical functions used for signal and image analysis. They are particularly well-suited for
representing signals with sharp transitions and discontinuities, making them useful for image compression,
denoising, and feature extraction. Wavelet transforms decompose an image into a set of wavelet coefficients
at different scales and orientations, providing a multi-scale representation of the image that can be used for
various image processing tasks.

You might also like