How can I use Principal Component Analysis (PCA) to reduce features?

4 views (last 30 days)
Hello
I extract image features using the Gabor filter.
The number of features is large for each image (5670 X 1) row and single column. How can I use Principal Component Analysis to reduce features?

Accepted Answer

Image Analyst
Image Analyst on 28 Sep 2018
Call PCA and then only keep the PC's that you want, like 5 or 10 of them. So, how does a Gabor filter give 5670 features? Are you saying each pixel is a feature?
  2 Comments
wisam kh
wisam kh on 30 Sep 2018
Thanks for your cooperation
gabor filter work on 8 orientations, so it give me large number of features.
suppose I have 40 features,
and call pca to reduce them, I got this error
Image Analyst
Image Analyst on 30 Sep 2018
imgaborfilt() gives you two images from every image you give it. Where are your 5670 or 40 different features coming from if they're not the individual pixels in either the magnitude or phase Gabor images?
You can't run pca on just one 1-d vector. It doesn't make sense. If you have 40 features, then you'd need those 40 measurements from at least 40 different images to get PCs.

Sign in to comment.

More Answers (0)

Categories

Find more on Dimensionality Reduction and Feature Extraction in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!